Are Suspects Who Resist Arrest Defiant,Desperate,or Disoriented?

Abstract

Objectives:

We examine three reasons why suspects resist arrest: (1) defiance of police authority by suspects from lower-status groups, (2) risky decisions resulting from aversion to sure losses, and (3) impairment due to mental illness and substance use.

Methods:

We use nationally representative survey data from about 17,000 state and federal inmates who were asked whether they resisted arrest when they committed the crime that led to their incarceration.

Results:

Suspects’ resistance is unrelated to their race/ethnicity, education, or unemployment. On the other hand, suspects are more resistant when they are carrying contraband (e.g., illegal weapons, drugs, stolen property) or are under community supervision (i.e., parolees, probationers, or escapees). Resistance is also positively related to mental illness, illicit drug use, and alcohol intoxication.

Conclusions:

Our results do not support the idea that resistance is an expression of defiance from lower-status suspects. They are consistent with prospect theory, which argues that decision makers become risk-seeking, when the alternative is to accept a sure loss. Our results suggest that resistant suspects are best understood as either desperate or disoriented decision makers.

Keywords

resisting arrest loss aversion mental illness prospect theory police

Recent events in Baltimore, Maryland and Ferguson, Missouri highlight the dangerous nature of arrest situations. Suspects who resist arrest expose themselves to injury and sometimes fatal violence by police officers (Skolnick and Fyfe 1993). In addition, police officers are at danger of being killed or injured, when they are making an arrest (Brandl 1996). In 2013, about 10 percent of police officers experienced an assault (U.S. Department of Justice 2014). Finally, resisting arrest may have collateral consequences for the general public. For example, between 1994 and 1998, high-speed pursuits in automobiles resulted in the death each year of an average of 114 bystanders (as well as 238 suspects; Hill 2002; see also Jacobs and Cherbonneau 2014).

Despite its importance, resisting arrest has generated only a limited amount of research. It appears in ethnographies of police culture (Brown 1981; Chevigny 1969; Muir 1977; Van Maanen 1978), as a control variable in studies of police use of force (Morabito et al. 2012; Terrill 2003; Terrill and Mastrofski 2002), and as a precipitating factor in research on the effects of the suspect’s demeanor on arrest (Mastrofski, Worden, and Snipes 1995; Worden and Shepard 1996; see also Mastrofski, Reisig, and McCluskey 2002). Most of this research has used the study of suspect resistance as a means to better understand police behavior, rather than as an object of inquiry itself.

Those studies that have focused on resistance to arrest or police authority have relied on data obtained from either the police or third-party observers (Belvedere, Worrall, and Tibbetts 2005; Bierie 2015; Bierie, Detar, and Craun 2013; Covington, Huff-Corzine, and Corzine 2014; Crawford and Burns 2002; Engel 2003; Greenleaf and Lanza-Kaduce 1995; Kavanagh 1997; Lanza-Kaduce and Greenleaf 2000; Mastrofski, Snipes, and Supina 1996; McCluskey, Mastrofski, and Parks 1999; Terrill 2003; Weidner and Terrill 2005).¹ These data sets have only limited information about suspects. This is a significant limitation, as a review of the evidence suggests the suspect characteristics that have been studied are better predictors of resistance than officer characteristics.

In the present study, we address this gap in the literature by examining self-report data from suspects. We use data from over 17,000 state and federal inmates who report whether they engaged in various forms of resistance during their most recent arrest. These forms range from legal oppositional behavior (e.g., arguing) to physical violence and evasive action. The data set has extensive information on the inmates that is not available in studies based on the reports of police or observers. Importantly, the data set includes measures of the suspect’s prior criminal behavior. Controlling for prior behavior reduces the likelihood that the relationships we observe are spurious.

Explanations of Resistance

We approach suspect resistance from three perspectives: as an expression of defiance from lower-status suspects; as a risky decision by suspects desperate to avoid prison; and as a function of the suspect’s impairment due to mental illness, drug use, and alcohol intoxication. We discuss each of the explanations in turn.

Resistance and Social Status

Many studies of why suspects resist arrest have emphasized the role of social inequality and the suspect’s location in the stratification system (e.g., Belvedere et al. 2005; Engel 2003). According to Black (1976, 1980), the application of the law is unequally distributed across social space. The availability of law as a resource is positively related to a citizen’s social status, while the application of law is negatively related to social status. Racial minorities and citizens with low socioeconomic status (SES) are less likely to enjoy the benefits of law enforcement when they are victimized, but they are more likely to suffer its consequences when they offend. As a result of mistreatment and discrimination, lower-status groups develop grievances that sometimes erupt into acts of violence against police officers. From Black’s (1980:36) perspective, many acts of violence against the police are actually “expressions of social control in retaliation for the misconduct of other officers.”

Tyler (1990) makes a similar argument but focuses on disrespectful treatment by police. He suggests that belief in the legitimacy of the police is strongly influenced by “procedural justice,” or whether citizens believe that authority figures have treated them in a fair, respectful manner.

When citizens experience police misconduct, they develop cynical attitudes toward law enforcement (Tyler 1997). In addition, citizens are sensitive to how treatment relates to their group membership and status, and these concerns influence whether they comply with authority (Tyler and Blader 2000). Sherman’s (1993) defiance theory extends Tyler’s work by arguing that police mistreatment may elicit retaliatory crime by defiant, resentful offenders who are alienated from conventional society. All of these theories imply that members of lower-status groups will be more likely to resist arrest.

Past research indicates that African Americans have more negative opinions of police officers than do other groups (Anderson 1999:34; Brunson 2007; Kochel, Wilson, and Mastrofski 2011:498; Tyler 2011). However, evidence for race differences in resistance to arrest is mixed. Bierie (2015) found that police encounters with Black offenders were more likely to result in violence against police than encounters with White or Hispanic suspects. Belvedere et al. (2005) found that African American suspects were significantly more likely than White or Hispanic offenders to physically resist arrest. Engel (2003) found that non-White suspects were significantly more likely to engage in noncompliant behavior than White suspects, when the arresting officer was White. However, no race differences were observed for verbal or physical aggression. In addition, the suspect’s race was not a significant predictor of resistance in other studies (Bierie et al. 2013; Covington et al. 2014; Crawford and Burns 2002; Greenlead and Lanza-Kaduce 1995; McCluskey et al. 1999; Weidner and Terrill 2005). Finally, Mastrofski et al. (1996) found that minority citizens were significantly more likely than White citizens to comply with the demands of White officers.

The lack of research on the effect of the suspect’s SES is probably due to the fact that direct information on SES is not available to police and third-party observers. Two studies, however, attempted to measure SES based on whether the suspect appeared to be poor. Mastrofski et al. (1996) found that this measure was negatively associated with whether the suspect complied with police requests, while McCluskey et al. (1999) found no relationship. No one has studied effects of SES using a self-report measure.

Resistance and Loss Aversion

From a rational choice perspective, suspects resist arrest when they believe the benefits are likely to outweigh the costs. They resist when they think they will be successful, particularly when they anticipate serious punishment if they comply. According to the “bounded” rationality model, suspects make reasonable decisions based on the information that is available to them (e.g., Cornish and Clarke 1986; March and Simon 1958).

It seems unlikely, however, that a strict rational choice model—even one involving bounded rationality—can fully account for why suspects resist arrest (Bierie et al. 2013; Mastrofski et al. 1996; McCluskey et al. 1999). First, evidence suggests that resistance has a low chance of success (Crew, Fridell, and Pursell 1995; Fennessy and Joscelyn 1971). This is not surprising, since suspects are committing an offense in the presence of police officers. Those officers are professionally trained to deny suspects the opportunity to resist arrest and to handle resistance when it occurs. In addition, the costs are substantial if the suspect fails. Suspects can anticipate more severe punishment, and there is the possibility they may be injured or killed by the police. Thus, suspects who resist are pursuing a line of action that does not have much chance of success and that is likely to be costly if it fails. It appears that they are only “making a bad situation worse.”

A strict cost–benefit analysis seems particularly problematic in understanding resisters in our sample of inmates, since they are reporting an arrest in which their resistance failed. Those who successfully resisted arrest could not be included in the sample. Of course, it may be that the resisters had a reasonable chance of success but were unlucky. However, it seems clear that most of them engaged in a behavior with a low probability of success and a costly outcome. It is difficult to interpret their act of futility as a rational choice. Insight from the field of behavioral economics, however, suggests that a modified rational choice approach that acknowledges the influence of cognitive biases may better explain why some suspects resist. The most well-established cognitive bias concerns loss aversion, whereby people are more heavily influenced by their fear of loss than by their hope for gain (Camerer 2004, 2005; Rabin 2000). The concept of loss aversion is a key tenet of prospect theory (Kahneman and Tversky 1979; Tversky and Kahneman 1992). According to prospect theory, loss aversion leads people to greater risk-taking when choosing between a guaranteed loss and a significantly larger loss that is merely probable. Kahneman (2011:319) describes this tendency as follows:

This is where people who face very bad options take desperate gambles, accepting a high probability of making things worse in exchange for a small hope of avoiding a large loss. Risk taking of this kind often turns manageable failures into disasters. The thought of accepting the large sure loss is too painful, and the hope of complete relief too enticing, to make the sensible decision that it is time to cut one’s losses.

The arrest situation seems like an ideal situation for the application of prospect theory. The theory offers an explanation for why suspects facing arrest would risk more severe punishment for a slim chance at escape. It suggests that people become risk-seeking—they go for “long shots”—when they are backed into a corner.

The effects of loss aversion on risk taking should be particularly strong in circumstances, where suspects face almost guaranteed incarceration if they are arrested. One circumstance is when suspects have been caught in the act, that is, in flagrante delicto. Another is the presence of contraband and other incriminating physical evidence at the scene. Finally, when suspects have a vulnerable legal status—they are escapees or on parole or probation—arrest virtually assures incarceration.

Only a few studies have examined these types of situations. Craun, Detar, and Bierie (2013) found that suspects with outstanding warrants were more likely to resist arrest with firearm violence. The strongest effects were observed for suspects with warrants for homicide and flight from justice. In a qualitative analysis, Margarita (1980) found that some offenders who killed police officers had panicked during routine traffic checks because they had incriminating evidence in their possession. Finally, there is research on the relationship between resistance and the quality of evidence against the suspect. One study found that a quality of evidence scale was positively associated with resistance (Weidner and Terrill 2005). However, McCluskey et al. (1999) found no effect, and Mastrofski et al. (1996) found an effect in the opposite direction. Mastrofski et al. claimed that the presence of evidence increased the suspect’s perception of the arresting officers’ legitimacy and therefore reduced suspect resistance.²

Only one of the studies described above attempted to control for prior criminal history. Craun et al. (2013) used a rough measure based on whether individual police officers believed the offender qualified as a “violent offender” or had a “significant criminal history.” It was not clear how they made this decision. It is important to control for prior record to reduce the likelihood of a spurious relationship. For example, serious offenders may be more likely to resist arrest and to have contraband. Research has examined whether suspects facing more serious charges are more likely to resist arrest (Belvedere et al. 2005; Bierie et al. 2013; Craun et al. 2013; Engel 2003; Greenleaf and Landza-Kaduce 1995; Kavanagh 1997; Mastrofski et al. 1996; McCluskey et al. 1999). The evidence is mixed.

Resistance and Cognitive Impairment

It is also possible that some suspects resist arrest as a result of impaired judgment. Suspects who are under the influence of alcohol or drugs, or who suffer from mental illness may be insensitive to costs, less attentive to social cues, or have unrealistic or delusional expectations. Muir (1977:127) observed that police officers were particularly concerned about dealing with “irrational persons—drug addicts, drunks, infuriated citizens, the deranged.”

Most of the research in this area focuses on alcohol intoxication. These studies consistently show that suspects who appear intoxicated are more likely to resist arrest (Bierie 2015; Bierie et al. 2013; Covington et al. 2014; Crawford and Burns 2002; Engel 2003; Kavanagh 1997). We are aware of only one study that has examined drug intoxication and resistance, and it found a positive relationship (Crawford and Burns 2002). It is unclear exactly how drug intoxication was measured in this study.

There have also been studies showing that suspects who resist arrest have higher scores on an irrationality index. The index was based on whether the observers thought the suspect showed signs of intoxication, mental illness, or heightened emotional arousal (Greenleaf and Lanza-Kaduce 1995; Lanza-Kaduce and Greenleaf 2000; Mastrofski et al. 1996; McCluskey et al. 1999; Weidner and Terrill 2005). Note that none of these studies of impairment and resistance control for the suspect’s prior record. It is possible that serious offenders are more likely to resist and are more likely to use drugs and alcohol.

No one, to our knowledge, has studied the effect of mental illness on resistance. There is, however, a substantial body of research linking mental illness with violent behavior (e.g., Engel and Silver 2001; Felson, Silver, and Remster 2012; Wallace, Mullen, and Burgess 2004). The relationship remains after controlling for substance abuse, which is associated with mental illness (e.g., Silver, Felson, and VanEseltine 2008).

Current Study

In this study, we examine predictors of resistance to arrest using data from a nationally representative sample of inmates in state and federal prisons. Our dependent variable is a variety scale comprised of nine questions regarding the respondents’ resistant behavior during the arrest that led to their current incarceration. In supplementary analyses, we examine whether the effects are different for different types of resistance.

Our analysis explores resistance from three different perspectives. First, we investigate whether suspects’ race and SES influence their level of resistance. Sociological theories of police–citizen conflict suggest that individuals with lower social status are more defiant during arrests because of their grievances with police. Second, we examine whether suspects carrying contraband or who are under community supervision (e.g., parolees, probationers, and escapees) are more likely to resist arrest. Both a rational choice approach and prospect theory predict that these suspects will be more resistant because they anticipate certain punishment if arrested. We emphasize the latter explanation because resistance has a low probability of success and high cost when it fails, and because the resisters in our sample failed.

Finally, we examine whether suspects under the influence of drugs or alcohol or who have a history of mental illness engage in more resistance. Impaired suspects may be more resistant because their ability to make rational decisions has been disrupted, or because their decisions are based on information that is unrealistic.

We include controls in our regression equations for each respondent’s prior criminal behavior, including their arrest history and the type of offense that led to their incarceration. These controls add a quasilongitudinal aspect to the design. It is important to control for past criminal behavior because it is likely that suspects who resist arrest are more serious offenders than suspects who do not resist.

We also include controls for the gender and age of the suspect. One might expect gender and age differences in resistance, given that men commit more crime than women, and offending declines as people get older (e.g., U.S. Department of Justice 2009). Bierie (2015) found that female offenders were less likely to assault police officers than male offenders. However, the majority of studies on resistance have either found that gender is unrelated to resistance, or that females are more likely to resist arrest (Covington et al. 2014; Mastrofski et al. 1996; McCluskey et al. 1999). It is surprising that most studies do not find effects of age (e.g., Belvedere et al. 2005; Covington et al. 2014; Mastrofski et al. 1996). Perhaps that is because age is measured by the estimates of observers. Two studies have found the predicted age effects. McCluskey et al. (1999) found that suspects who appeared to be under the age of 21 were more likely to resist arrest. In addition, Bierie (2015) found that police encounters with offenders were more likely to result in an officer being assaulted, when the police report indicated that at least one of the offenders was a juvenile.

Data and Methods

The data were based on the 2004 Survey of State and Federal Correctional Facilities (U.S. Department of Justice 2007). The sample is nationally representative of the U.S. inmate population in that year. Computer-assisted personal interviews were conducted at 287 state prisons and 39 federal prisons. Stratified random sampling was used to select prison facilities based on the size of the facility’s inmate population relative to the national inmate population, census region, state population, and whether the prison housed male or female inmates. The original sample included 18,185 cases (14,499 state inmates and 3,686 federal inmates). The survey researchers relied on estimates from the Census of State and Federal Correctional facilities to inform their sampling strategy. After collecting the data, the research team constructed survey weights to adjust for the sampling procedure, such as the under sampling of drug offenders (for more details, see U.S. Department of Justice 2007). To account for the complex survey design, we applied survey weights to the data using Stata 12’s svy command (StataCorp 2009a). We dropped information from 858 inmates because they turned themselves in to the police or were already incarcerated at the time of their arrest. In an additional 61 cases, inmates provided values for their age at arrest that were too young to be credible (discussed below), so these inmates were excluded from analyses; consequently, our final sample fell to 17,266 inmates.

Our dependent variable is a variety scale comprised of nine questions regarding the respondents’ behavior during the arrest that led to their current incarceration. Respondents were asked whether they engaged in any of the following behaviors at any time during the arrest: “did you argue with or disobey the police officer(s)?”; “curse at, insult, or call the police officer(s) a name?”; “say something threatening to the police officer(s)?”; “resist being handcuffed or arrested?”; “resist being searched or having your vehicle searched?”; “try to escape by hiding, running away, or engaging in a high-speed chase?”; “grab, hit, or fight with the police officer(s)?”; “use a weapon to threaten the police officer(s)?”; and “use a weapon to assault the police officer(s)?” To construct the variety scale, respondents received a score of 1 for each item that they responded “yes” to; these scores were then summed. The final scale ranges from 0 to 9, with higher scores indicating a greater diversity and level of resistant behavior.

By constructing our dependent variable as a variety scale, we can treat our outcome measure as a count of how many types of resistant behaviors each respondent reported. This allows us to use negative binomial regression, which is equipped to deal with overdispersion in the dependent variable that arises from a large proportion of zero counts (Long 1997; MacDonald and Lattimore 2010; Osgood 2000). For a comprehensive review of variety scales, see Sweeten (2012).

An apparent limitation of the scale is that acts of resistance are given the same weight regardless of their seriousness. However, since serious acts are much less frequent than minor ones, suspects who engage in a serious form of resistance usually also engaged in more minor forms. Analyses of our data confirmed this pattern. For example, few respondents (n = 15) who engaged in or threatened to engage in violence committed only a single act of resistance.

We created four dummy variables to convey inmate race/ethnicity: non-Hispanic Black, Hispanic, non-Hispanic other race, and non-Hispanic White (the reference group). We include two measures of the inmates’ SES: educational attainment and legal employment during the month prior to arrest. Educational attainment refers to the highest year of school completed. We created a dummy variable to represent whether the respondent was employed or unemployed in the month preceding arrest.

Inmates were also asked whether they were on parole, on probation, or had escaped from police custody at the time of their last arrest. We created a series of dummy variables to represent each respondent’s criminal status, where those who did not have an active criminal status were placed in the reference group. Note that parolees are an important category because there are so many of them and because their numbers continue to increase. Lin (2010) estimates that 22.2 percent of U.S. inmates in 2008 were serving time due to parole revocations, compared to 7.1 percent of inmates in 1980 (see also Glaze and Parks 2012).

Inmates were also asked whether the police found illegal weapons, illegal drugs, stolen property, or other evidence of a crime. Based on their response to these questions, we created a series of contraband dummy variables. Respondents were coded 1 if they possessed the relevant item and 0 if they did not.

The measure of mental health was based on questions about treatment during the year before arrest. Inmates were asked whether they had experienced any of the following as a result of a “mental or emotional problem”: (1) taking medication prescribed by a doctor; (2) admitted to a mental hospital, unit, or treatment program where they stayed overnight; (3) received counseling or therapy from a trained professional; and (4) received any other mental health treatment or services. If the inmate answered “yes” to any of these questions, they were coded as having experienced mental health problems.

Inmates were also asked whether or not they were tested for drugs at the time of arrest, and if so, whether they tested positive. We created dummy variables based on whether they tested positive, they were not tested, or whether they tested negative (the reference category). Inmates were also asked how frequently they consumed alcohol during the year before their last arrest. We created four dummy variables to represent the respondents’ frequency of alcohol consumption: daily, weekly, less than weekly, and nondrinkers (the reference category). We also created a dummy variable based on whether the police found an open alcohol container on the respondent at the time of arrest. Unfortunately, we do not have a direct measure of alcohol intoxication during the incident.

Our equations included controls for gender, age, and prior criminal behavior. We computed the inmates’ age at the time of their last arrest by subtracting the year of arrest from the year of the survey and then subtracting this value from their age at the time of the survey.³ Prior criminal behavior was based on measures of arrest history and the nature of their current offense. Arrest history was based on the total number of prior arrests reported by the respondent. Respondents who reported 10 or more prior arrests were assigned a code of 10 based on our concern that offenders who had been arrested more than 10 times would have difficulty recalling the precise number. We also included a quadratic term for arrest record, since preliminary analyses indicated curvilinearity. Current offense was coded according to the main offense that led to their current incarceration. In instances where offenders were serving time for multiple offenses, the inmate was classified based on the offense with the longest sentence. This variable included four general categories: violent offenses (e.g., murder, assault, rape, robbery), property offenses (e.g., theft, burglary, arson), public order offenses (e.g., public intoxication, bribery, contempt of court), and drug offenses (e.g., possession of a controlled substance, drug trafficking, drug smuggling). Property offenses served as the reference group.

The majority of variables had less than 1 percent missing values. The greatest amount of missing values was found for the drug test measure (6 percent missing). To handle missing data, we multiply imputed 10 data sets using multivariate normal regression in Stata 12 (StataCorp 2009b, 2011). This procedure relies on a Markov chain Monte Carlo method to impute values that are missing at random (Rubin 2004). The pattern of results was the same regardless of whether we used multiply imputed data or listwise deletion.

Results

Descriptive statistics are presented in Table 1. The table shows that resisting arrest is relatively rare—around 15 percent of respondents engaged in some form of resistance. This frequency is similar to what is reported in studies based on police records and observational data; they report frequencies between 12 percent and 18 percent (Crawford and Burns 2002; Engel 2003; Garner et al. 1995; Leinfelt 2005; U.S. Department of Justice 2012).

Table 1.

Sample Descriptives.

	Variable	Mean/Percent	Std. Dev.
Resistance^a	None	84.7
	Argue/disobey	7.9
	Curse at/insult	5.7
	Threaten	1.3
	Resist handcuffs	3.7
	Resist search	1.7
	Try to escape	6.3
	Fight	1.1
	Threaten with weapon	0.3
	Assault with weapon	0.2
Resistance (variety scale)		$\bar{x} = 0.28$	0.83
Education		$\bar{x} = 10.5$	2.58
Employed		70.2
Race/ethnicity	Black	39.9
	Other race	6.2
	Hispanic	19.2
	White	34.7
Community supervision status	Probationer	23.3
	Parolee	16.7
	Escapee	0.5
	None	59.5
Contraband^a	Illegal weapons	7.1
	Illegal drugs	16.7
	Stolen property	6.5
	Other	5.1
	Nothing	51.6
Mental health problems		13.5
Drug test results	Positive	8.6
	Negative	16.4
	Not tested	75.0
Alcohol consumption	Daily	23.9
	Weekly	20.2
	>Weekly	24.0
	Nondrinker	31.9
Open alcohol container		3.0
Female		20.9
Age at arrest		$\bar{x} = 30.9$	9.99
No. of previous arrests		$\bar{x} = 3.7$	3.32
Most recent offense	Violent	45.3
	Drug	25.3
	Public order	6.9
	Property	22.5

Note: N = 17,266.

^aTotals do not add to 100, as respondents could report multiple forms of resistance or contraband.

In general, the distribution of resistant behaviors follows a pattern: the more serious the resistant behavior, the lower its frequency. For example, verbal aggression was much more frequent than physical violence, and the use of weapons was rare. Descriptive statistics for the variety scale show that overdispersion is an issue: The variable’s variance is much larger than its mean.

Roughly 40 percent of all respondents were under community supervision at the time of their last arrest. The most widespread status was being on probation (23.3 percent), while the least common was escape from police custody (0.5 percent). In addition, approximately 35 percent of the sample was carrying some form of contraband. The most common type of contraband was illegal drugs (16.7 percent). Finally, around 14 percent of respondents indicated that they had received treatment for mental health problems during the year before their last arrest.

We present the results from our negative binomial regression model in Table 2. The results do not support the hypothesis that defiance based on social status leads to resistance. Blacks and Hispanics were no more resistant than Whites. Resistance was also unrelated to education and employment status.

Table 2.

Results from Negative Binomial Regression Predicting Resistance Variety Score (N = 17,266).

Variable	Incident Risk Ratio
Social status
Education	0.98
Employed	0.92
Black	0.98
Other race	0.97
Hispanic	0.93
Loss aversion
Probationer	1.16**
Parolee	1.27***
Escapee	1.63
Weapons	1.45***
Drugs	1.30***
Stolen property	1.60***
Other evidence	1.36***
Impaired judgment
Mental health problems	1.52***
Positive drug test	1.46**
No drug test	1.02
Daily alcohol	1.27***
Weekly alcohol	1.13
>Weekly alcohol	1.04
Open alcohol container	1.46**
Control variables
Female	0.84*
Age at arrest	0.96***
Number of arrests	1.16***
Number of arrests²	0.99*
Violent offense	1.16
Drug offense	1.03
Public order offense	1.23

*p < .05.

**p < .01.

***p < .001.

We do find evidence for the loss aversion hypothesis. The table shows that respondents who were under community supervision or carrying contraband at the time of arrest engaged in a significantly greater variety of resistant behaviors. For instance, suspects who were on parole at the time of their last arrest committed roughly 27 percent [Incident Risk Ratio (I.R.R.) = 1.27] more distinct acts of resistance than suspects who were not under community supervision. The effect for escapees (I.R.R. = 1.63) is substantial, but did not achieve statistical significance, probably due to the small number of escapees (n = 86). All of our measures of contraband are associated with an increased number of acts of resistance. For example, suspects who were caught with stolen property engaged in about 60 percent more resistant behaviors than suspects who were not carrying stolen property.

The findings from Table 2 also support the hypothesis that impairment leads to more resistance. Respondents who received treatment for a mental health issue report more resistant acts than those who had not received such treatment. Suspects with mental health issues showed roughly 52 percent more resistant behaviors. Suspects who tested positive for drug use at the time of arrest were more resistant than suspects who reported a negative drug test. We also observed effects of alcohol. Suspects were more likely to engage in a greater variety of resistant acts if they drank on a daily basis or if they were carrying an open alcohol container.

The table also shows that suspects who were male and younger were more resistant than their counterparts. Finally, arrest record was positively associated with resistance. However, the significant squared term indicates the relationship is curvilinear. The fact that the effect of arrest record tapers off suggests that once an offender has a sufficient number of arrests, an additional arrest is not indicative of greater criminality.

Supplementary Analyses

In supplementary analyses, we examined whether the results would be different if we examined specific types of resistance using multinomial logistic regression (see Appendix). We created the following outcomes: verbal resistance, evasive resistance, violent resistance, and no resistance (the reference category) based on the questions described in the methods section. Those respondents who engaged in multiple types of resistance were coded according to their most serious act of resistance. For example, respondents were coded as having engaged in violent resistance if they engaged in both violence and evasive action. The results are consistent with the results from the negative binomial model. We find no evidence that lower-status suspects are more likely to resist arrest. In fact, Hispanic suspects are significantly less likely than White suspects to have engaged in verbal resistance. The effects of loss aversion and impairment are primarily observed for evasive and violent forms of resistance. Not all effects are statistically significant, but the pattern is clear.

We also estimated a zero-inflated negative binomial model as a sensitivity analysis for our negative binomial model with the resistance variety score. Some scholars prefer this method when there are a large number of zeroes in the dependent variable (e.g., Lambert 1992; Mullahy 1986). This estimation procedure adjusts for the excess zeroes by modeling two separate processes: a binary model where the outcome is the probability of having a count value of zero and a model for estimating the expected count value (Long 1997). We included the same set of covariates in the zero-inflated model (both the binary model and the count model) that we used for our main analysis. The results from the zero-inflated model were similar to the results we present.

In other analyses not shown, we examined whether memory issues could be affecting our results. Most inmates were describing fairly recent arrests, but for some the arrests occurred many years prior to the interview. We added to our equation a continuous measure of the time lag between the arrest encounter and the date of the inmate interview. The coefficient was close to zero and its inclusion did not affect the results.

In an additional analysis, we substituted a measure of mental illness based on whether the inmate had ever been officially diagnosed. The effect was similar to the effect of treatment. We also looked for statistical interactions between age and race, age and sex, sex and race, and age/sex/race. However, we did not find any significant multiplicative effects.

Finally, we performed further analyses to see if we could detect any effects of the status variables. We examined bivariate relationships between the status variables and resistance and found no effects. This suggests that we are not masking effects of suspect race, unemployment, and education by controlling for other variables. We examined the effects of occupational prestige for employed inmates by using scores from the Stevens and Cho (1985) socioeconomic index. Occupational prestige was unrelated to resistance.

Discussion

This is the first study to use self-report data from offenders to examine resisting arrest. The use of self-report data allowed us to better measure some key variables and to control for prior criminal behavior. We focused on three sources of resistant behavior: (1) defiance of police authority by suspects from lower-status groups, (2) risky decisions resulting from aversion to sure losses, and (3) impairment due to mental illness and substance use. We discuss each of these sources in turn.

Defiance

Our results do not support the idea that resisting arrest reflects the social status of suspects. African Americans are no more resistant (or likely to resist) than Whites. This result is consistent with most prior research. One could argue that Black suspects are only defiant, when the arresting officer is White (e.g., Engel 2003). We did not have a measure of the officer’s race, so we could not examine race by race statistical interactions. However, the majority of officers are White (Reaves 2011), and we should have detected at least a weak main effect of race if Blacks were defiant against White officers. The coefficient for race was close to zero and in the opposite direction predicted.

It seems more likely that any tendency for Black suspects to be defiant is offset by fear. African Americans may be more likely than Whites to anticipate harsh treatment if they resist. From a rational choice perspective, Blacks who feel threatened by police should be more compliant when confronting them, not less. Future research should measure the suspect’s attitudes toward police and their beliefs about the dangerousness of resistance in order to disentangle these effects. Note that the relationship between race and resistance may have changed recently due to Black protests about police violence.

We found no evidence suggesting that other indicators of low social status affected resistance. Hispanics were no more resistant to arrest than Whites and were less likely to argue with or insult police. Suspects who lacked education or who were unemployed were no more resistant than their counterparts. Our results do not support the idea that inequality leads suspects to be defiant and resist arrest.

It may be, however, that some suspects resist arrest because they are defiant for other reasons. Membership in lower-status groups is not the only factor that might lead to defiance. For example, it may be that alienation from conventional society leads suspects to resist arrest (Sherman 1993), regardless of their social status. Perhaps innocent suspects are more defiant, when they are arrested. According to reactance theory, people are likely to resist social pressures when they believe their freedom is being unfairly constrained, even when resistance is costly (Brehm and Brehm 2013). Experimental research shows that participants sometimes do the opposite of what they are told to do when they think that the influence is illegitimate (e.g., Fogarty 1997; Miron and Brehm 2006; Quick and Bates 2010).

Suspects may also be more defiant if they believe the officer confronting them has been disrespectful or has been too aggressive in making the arrest (e.g., Brunson 2007; McCluskey 2003; Watson and Angell 2013). Issues of procedural justice and interactional justice may come into play (e.g., Mikula, Petrik, and Tanzer 1990).

Loss Aversion

We found that suspects were more resistant if they were on probation or parole or were carrying any type of contraband. From a rational choice perspective, one would expect more resistance in these circumstance since an arrest means certain punishment. On the other hand, a rational choice explanation might be questioned since resistance rarely succeeds, and when it fails the costs are substantial. In addition, the resisters in our sample did in fact fail. While their apprehension could be attributed to bad luck—we are judging them in hindsight—it seems likely that many of them made a bad choice.

Because of these considerations, we suggested that prospect theory could contribute to understanding resistance. Arrest encounters involving contraband and vulnerable legal statuses are precisely the types of situations where the cognitive biases identified by prospect theory are expected to be strongest. Loss aversion increases the likelihood that suspects will resist in spite of the low probability of success and the high risk of negative outcomes. In addition, there is clear experimental evidence supporting the effects of loss aversion on risk taking (e.g., Abdellaoui, Bleichrodt, and L’Haridon 2008; Curley, Yates, and Abrams 1986; Hogarth and Einhorn 1990; Thaler et al. 1997). The ideas are widely accepted, as indicated by the awarding of a Nobel Prize to Kahneman for his original work in this area (Kahneman and Tversky 1979). Criminologists have been borrowing these ideas, particularly in the understanding of deterrence (e.g., Bushway and Owens 2013; Cornish and Clarke 2014; Loughran, Paternoster, and Weiss 2012).

Other Applications of Loss Aversion

We believe that loss aversion can contribute to an understanding of other crimes that seem to have an extremely low probability of success or that seem particularly cruel or “senseless.” Offenders who have committed such crimes may have been intoxicated, mentally ill, or lacking in intelligence, but they may also have been attempting to avoid a sure loss.

For example, loss aversion suggests that the tendency to retaliate when insulted should be stronger than a traditional rational choice approach would suggest. Targets of verbal attacks sometimes risk serious injury if they retaliate but they will inevitably lose face if they do nothing. Antagonists may also be more likely to use guns and other weapons when they are losing than one would predict from a simple cost–benefit analysis. The risk of severe punishment increases, but they take the risk in order to avoid a sure loss.

Loss aversion may also increase the likelihood of lethal intent and homicide in certain circumstances. Offenders are likely to anticipate a certain loss if they know that the victim or other witnesses can identify them. When offenders believe that allowing a witness to survive will lead to incarceration, homicide is more likely than one would expect from a strict cost–benefit analysis. The perception of certain loss is also likely to lead to lethal intent when people are in conflict with an armed or otherwise dangerous adversary. They are likely to take great risks when it is “kill or be killed” and may engage in preemptive strikes when they anticipate an attack.

Impairment

We found effects of impairment due to alcohol, illicit drug use, and mental illness. The effects of alcohol are consistent with prior research. Suspects were more resistant if they drank on a daily basis or if they were carrying an open alcohol container. It is noteworthy that we observe an effect of alcohol intoxication even with a weak measure.

We were more interested in the effects of illicit drugs. We found that suspects who tested positive for illicit drug use exhibited higher levels of resistance. Our results are consistent with Crawford and Burns (2002) who measured drug use based on whether the arresting officer reported that the suspect was under the influence. We could not determine what types of drugs were producing the effects.

We also find that suspects who had been treated for mental illness reported more acts of resistance than suspects who had not been treated and were more likely to violently resist arrest. These results are consistent with studies cited earlier that find a relationship between mental illness and violence in other circumstances. The effects of mental illness cannot be attributed to the effects of alcohol intoxication, as we controlled for both individual differences in the frequency of drinking and the presence of open alcohol containers at the time of arrest. However, we acknowledge the limitations of our measure of mental illness. Undoubtedly, only certain forms of mental illness are related to resistance and other aggressive behavior (e.g., Felson et al. 2012; Link, Andrews, and Cullen 1992). Future research should rely on better measurement to examine these issues.

The relationship between mental illness and resisting arrest has policy implications. Beginning with the deinstitutionalization of state mental hospitals during the 1960s, the criminal justice system has gradually become the nation’s largest provider of mental health care services (Harcourt 2011; Markowitz 2006; Torrey et al. 2010). In the current era of community policing, citizens’ mental health problems routinely become police problems (Cordner 2006). As a result, police officers have become increasingly involved in managing the crises of people with mental illness (Lamb, Weinberger, and Gross 2004). Our analysis suggests that police departments should prioritize training programs to help prepare officers for this task (e.g., Reuland, Schwarzfeld, and Draper 2009).

Limitations

To perform our analyses, it was necessary to rely on the self-reports of inmates about events that could have occurred in the distant past. Some inmates may have forgotten whether they resisted, whether they had contraband, or whether they were on probation or parole at the time. Their memory of events could be clouded if they were under the influence of drugs or alcohol at the time. On the other hand, it could be argued that arrests are salient events in people’s lives that are unlikely to be forgotten. Moreover, the data used in this study were collected using face-to-face interviews, which tend to reduce memory effects by allowing respondents sufficient time for memory retrieval (Junger-Tas and Marshall 1999). Nevertheless, we attempted to address the recall issue by controlling for the recency of the arrest. Recency was not associated with resistance and the effect of our other variables on resistance did not depend on the recency of the arrest. We therefore are doubtful that measurement error due to memory failure is creating serious problems.

We suspect that what measurement error exists in our variables is probably random and not systematic. Measurement error in the independent variables (but not the dependent variable) is likely to bias coefficients downward (Allison 1999). Perhaps the effects we found would have been stronger if we had better measurement. Perhaps we would have obtained effects, where we did not. Note, however, that the variables of interest that did not have effects were race, education, and employment status; it is unlikely that these variables have much measurement error.

Conclusion

Rewards and costs affect the decision to resist arrest but so do other factors. Our research suggests that impairment resulting from cognitive biases, drug and alcohol use, and mental illness increase the likelihood of resistance. No doubt some resistance is an act of defiance, but we suspect it more likely to be a reaction to the behavior of the arresting officer than attitudes toward the legal system. We also suspect that the negative attitudes of African Americans toward the police are just as likely to lead to fear and compliance as they are to lead to defiance.

We interpret our results as pointing to the importance of situational factors in resistance. Possession of contraband, for example, is a situational factor that can fluctuate over a short time span. A man who is pulled over by the police may behave very differently depending on whether he is driving to or from his drug dealer. On the return trip he is carrying contraband, faces certain punishment, and is therefore more likely to resist arrest. When suspects find themselves in hopeless arrest situations, cognitive biases incline them toward resistance. The certainty of loss eclipses other considerations, and high-risk behavior becomes more appealing.

Footnotes

Appendix

Table A1.

Odds Ratios from Multinomial Logistic Regression Predicting Resistance Types versus No Resistance (N = 17,266).

Variable	Verbal	Evasive	Violent
Social status
Education	1.03	0.83	0.97
Employed	0.91	0.92	0.92
Black	0.99	1.21	1.03
Other race	0.99	1.27	0.90
Hispanic	0.73***	1.04	1.03
Loss aversion
Probationer	1.20	1.15	1.17
Parolee	1.07	1.72***	1.27*
Escapee	0.89	2.34	1.93
Weapons	1.08	1.29*	1.34*
Drugs	1.16	1.27	1.49**
Stolen property	1.45*	3.13***	1.88***
Other evidence	1.45*	1.52*	1.57**
Impaired judgment
Mental health problems	1.19	1.07	1.63***
Positive drug test	1.27	1.80**	1.67**
No drug test	1.05	1.39*	1.03
Daily alcohol	1.17	1.14	1.23*
Weekly alcohol	1.26*	1.16	1.05
>Weekly alcohol	1.04	1.06	1.02
Open alcohol container	1.63**	1.21	1.25
Control variables
Female	0.84	0.51***	0.66*
Age at arrest	0.96***	0.96***	0.96***
Number of arrests	1.25***	1.32***	1.14**
Number of arrests²	0.99**	0.98***	0.99
Violent offense	0.95	1.09	1.54***
Drug offense	1.34**	0.84	0.99
Public order offense	1.09	1.51**	1.55*

*p < .05.

**p < .01.

***p < .001.

Authors’ Note

The authors contributed equally to the paper.

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

References

Abdellaoui

Bleichrodt

l’Haridon

. 2008. “A Tractable Method to Measure Utility and Loss Aversion under Prospect Theory.” Journal of Risk and Uncertainty 36:245–66.

Allison

Paul D.

1999. Multiple Regression: A Primer. Thousand Oaks, CA: Pine Forge Press.

Anderson

1999. Code of the Street. New York: Norton.

Belvedere

Worrall

J. L.

Tibbetts

S. G.

. 2005. “Explaining Suspect Resistance in Police-citizen Encounters.” Criminal Justice Review 30:30–44.

Bierie

David M.

2015. “Assault of Police.” Crime & Delinquency. doi: 10.1177/0011128715574977.

Bierie

D. M.

Detar

P. L.

Craun

S. W.

. 2013. “Firearm Violence Directed at Police.” Crime & Delinquency. 0011128713498330.

Black

1976. The Behavior of Law. New York: Academic Press.

Black

1980. The Manners and Customs of the Police. New York: Academic Press.

Brandl

S. G.

1996. “In the Line of Duty: A Descriptive Analysis of Police Assaults and Accidents.” Journal of Criminal Justice 24:255–64.

10.

Brehm

S. S.

Brehm

J. W.

2013. Psychological reactance: A theory of freedom and control. Academic Press. New York.

11.

Brown

M. K.

1981. Working the Street: Police Discretion and the Limits of Reform. New York: Russell Sage Foundation.

12.

Brunson

R. K.

2007. ““Police Don’t Like Black People”: African‐American Young Men’s Accumulated Police Experiences*.” Criminology & Public Policy 6:71–101.

13.

Bushway

S. D.

Owens

E. G.

. 2013. “Framing Punishment: Incarceration, Recommended Sentences, and Recidivism.” Journal of Law and Economics 56:301–31.

14.

Camerer

C. F.

2004. “Prospect Theory in the Wild: Evidence from the Field.” Pp. 148–61 in Advances in Behavioral Economics, edited by Colin

Camerer

George Loewenstein

Matthew

Rabin

. Princeton, NJ: Princeton University Press.

15.

Camerer

C. F.

2005. “Three Cheers—Psychological, Theoretical, Empirical—for Loss Aversion.” Journal of Marketing Research 42:129–33.

16.

Chevigny

1969. Police Power: Police Abuses in New York City. New York: Pantheon books.

17.

Cornish

Clarke

. 1986. The Reasoning Criminal. New York: Springer-Verlag.

18.

Cornish

Clarke

, eds. 2014. The Reasoning Criminal: Rational Choice Perspectives on Offending. New Brunswick, NJ: Transaction.

19.

Cordner

2006. People with mental illness. Washington, DC: Office of Community Oriented Policing Services.

20.

Covington

M. W.

Huff-Corzine

Corzine

. 2014. “Battered Police: Risk Factors for Violence against Law Enforcement Officers.” Violence and Victims 29:34–52.

21.

Craun

Sarah W.

Detar

Paul J.

Bierie

David M.

. 2013. “Shots Fired: Firearm Discharges During Fugitive Apprehensions.” Victims & Offenders 8:56–69.

22.

Crawford

Burns

. 2002. “Resisting Arrest: Predictors of Suspect Non-compliance and Use of Force against Police Officers.” Police Practice and Research 3:105–17.

23.

Crew

R. E.

Fridell

L. A.

Pursell

. 1995. “Probabilities and Odds in Hot Pursuit: A Benefit-cost Analysis.” Journal of Criminal Justice 23:417–24.

24.

Curley

S. P.

Yates

J. F.

Abrams

R. A.

. 1986. “Psychological Sources of Ambiguity Avoidance.” Organizational Behavior and Human Decision Processes 38:230–56.

25.

Engel

R. S.

2003. “Explaining Suspects’ Resistance and Disrespect toward Police.” Journal of Criminal Justice 31:475–92.

26.

Engel

R. S.

Silver

. 2001. “Policing Mentally Disordered Suspects: A Reexamination of the Criminalization Hypothesis*.” Criminology 39:225–52.

27.

Felson

R. B.

Silver

Remster

. 2012. “Mental Disorder and Offending in Prison.” Criminal Justice and Behavior 39:125–43.

28.

Fennessy

E. F.

Jr Joscelyn

K. B.

. 1971. “National Study of Hot Pursuit, A.” Denver Law Journal 48:389.

29.

Fogarty

J. S.

1997. “Reactance Theory and Patient Noncompliance.” Social Science & Medicine 45:1277–88.

30.

Garner

J. H.

Schade

Hepburn

Buchanan

. 1995. “Measuring the Continuum of Force Used by and Against the Police.” Criminal Justice Review 20(2):146–168.

31.

Glaze

Parks

. 2012. Correctional Populations in the United States, 2011. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

32.

Greenleaf

R. G.

Lanza-Kaduce

. 1995. “Sophistication, Organization, and Authority-subject Conflict: Rediscovering and Unraveling Turk’s Theory of Norm Resistance*.” Criminology 33:565–86.

33.

Harcourt

B. E.

2011. “Reducing Mass Incarceration: Lessons from the Deinstitutionalization of Mental Hospitals in the 1960s.” University of Chicago Law & Economics, Olin Working Paper no. 542, Chicago.

34.

Hill

2002. “High-speed Police Pursuits: Dangers, Dynamics, and Risk Reduction.” FBI Law Enforcement Bulletin 71:14–18.

35.

Hogarth

R. M.

Einhorn

H. J.

. 1990. “Venture Theory: A Model of Decision Weights.” Management Science 36:780–803.

36.

Jacobs

B. A.

Cherbonneau

. 2014. “Auto Theft and Restrictive Deterrence.” Justice Quarterly 31:344–67.

37.

Junger-Tas

Josine

Marshall

Ineke Haen

. 1999. “The Self-report Methodology in Crime Research.” Crime and Justice 25:291–367.

38.

Kahneman

2011. Thinking, Fast and Slow. Straus and Giroux. New York.

39.

Kahneman

Tversky

. 1979. “Prospect Theory: An Analysis of Decision under Risk.” Econometrica: Journal of the Econometric Society 47:263–91.

40.

Kavanagh

1997. “The Occurrence of Resisting Arrest in Arrest Encounters: A Study of Police-citizen Violence.” Criminal Justice Review 22:16–33.

41.

Kochel

T. R.

Wilson

D. B.

Mastrofski

S. D.

. 2011. “Effect of Suspect Race on Officers’ Arrest Decisions*.” Criminology 49:473–512.

42.

Lamb

H. R.

Weinberger

L. E.

Gross

B. H.

. 2004. “Mentally Ill Persons in the Criminal Justice System: Some Perspectives.” Psychiatric Quarterly 75:107–26.

43.

Lambert

1992. “Zero-inflated Poisson Regression, with an Application to Defects in Manufacturing.” Technometrics 34:1–14.

44.

Lanza-Kaduce

Greenleaf

R. G.

. 2000. “Age and Race Deference Reversals: Extending Turk on Police-citizen Conflict.” Journal of Research in Crime and Delinquency 37:221–36.

45.

Leinfelt

F. H.

2005. “Predicting use of Non-lethal Force in a Mid-size Police Department: A Longitudinal Analysis of the Influence of Subject and Situational Variables.” The Police Journal 78(4):285–300.

46.

Lin

2010. “Parole Revocation in the Era of Mass Incarceration.” Sociology Compass 4:999–1010.

47.

Link

B. G.

Andrews

Cullen

F. T.

. 1992. “The Violent and Illegal Behavior of Mental Patients Reconsidered.” American Sociological Review 57:275–92.

48.

Long

J. S.

1997. Regression Models for Categorical and Limited Dependent Variables. Sage Publications. California.

49.

Loughran

T. A.

Paternoster

Weiss

. 2012. “Hyperbolic Time Discounting, Offender Time Preferences and Deterrence.” Journal of Quantitative Criminology 28:607–28.

50.

MacDonald

Lattimore

. 2010. “Count Models in Criminology.” Pp. 683–98 in Handbook of Quantitative Criminology, edited by Piquero

Weisburd

. Springer.

51.

March

J. G.

Simon

H. A.

. 1958. Cognitive Limits on Rationality. New York: Wiley.

52.

Margarita

1980. “Killing the Police: Myths and Motives.” The Annals of the American Academy of Political and Social Science 452:63–71.

53.

Markowitz

F. E.

2006. “Psychiatric Hospital Capacity, Homelessness, and Crime and Arrest Rates*.” Criminology 44:45–72.

54.

Mastrofski

S. D.

Reisig

M. D.

McCluskey

J. D.

. 2002. “Police Disrespect toward the Public: An Encounter-based Analysis*.” Criminology 40:519–52.

55.

Mastrofski

S. D.

Snipes

J. B.

Supina

A. E.

. 1996. “Compliance on Demand: The Public’s Response to Specific Police Requests.” Journal of Research in Crime and Delinquency 33:269–305.

56.

Mastrofski

S. D.

Worden

R. E.

Snipes

J. B.

. 1995. “Law Enforcement in a Time of Community Policing.” Criminology 33:539.

57.

McCluskey

J. D.

2003. Police Requests for Compliance: Coercive and Procedurally Just Tactics. New York: LFB Scholarly Pub.

58.

McCluskey

J. D.

Mastrofski

S. D.

Parks

R. B.

. 1999. “To Acquiesce or Rebel: Predicting Citizen Compliance with Police Requests.” Police Quarterly 2:389–416.

59.

Mikula

Petri

Tanzer

. 1990. “What People Regard as Unjust: Types and Structures of Everyday Experiences of Injustice.” European Journal of Social Psychology 20:133–49.

60.

Miron

A. M.

Brehm

J. W.

. 2006. “Reactance Theory-40 Years Later.” Zeitschrift für Sozialpsychologie 37:9–18.

61.

Morabito

M. S.

Kerr

A. N.

Watson

Draine

Ottati

Angell

. 2012. “Crisis Intervention Teams and People with Mental Illness Exploring the Factors that Influence the Use of Force.” Crime & Delinquency 58:57–77.

62.

Muir

W. K.

1977. Police: Streetcorner Politicians. Chicago: University of Chicago Press.

63.

Mullahy

1986. “Specification and Testing of Some Modified Count Data Models.” Journal of Econometrics 33:341–65.

64.

Osgood

D. W.

2000. “Poisson‐based Regression Analysis of Aggregate Crime Rates.” Journal of Quantitative Criminology 16:21–44.

65.

Quick

B. L.

Bates

B. R.

. 2010. “The Use of Gain-or Loss-frame Messages and Efficacy Appeals to Dissuade Excessive Alcohol Consumption among College Students: A Test of Psychological Reactance Theory.” Journal of Health Communication 15:603–28.

66.

Rabin

2000. Risk Aversion and Expected-utility Theory: A Calibration Theorem. Berkeley: Department of Economics, University of California at Berkeley.

67.

Reaves

B. A.

2011. Local Police Departments (2007). Darby, PA: DIANE.

68.

Reuland

Schwarzfeld

Draper

. 2009. Law Enforcement Responses to People with Mental Illnesses: A Guide to Research-informed Policy and Practice. New York: Council of State Governments Justice Center.

69.

Rubin

D. B.

2004. Multiple Imputation for Nonresponse in Surveys. Vol. 81. New York: John Wiley.

70.

Sherman

L. W.

1993. “Defiance, Deterrence, and Irrelevance: A Theory of the Criminal Sanction.” Journal of Research in Crime and Delinquency 30:445–73.

71.

Silver

Felson

R. B.

Vaneseltine

. 2008. “The Relationship between Mental Health Problems and Violence among Criminal Offenders.” Criminal Justice and Behavior 35:405–26.

72.

Skolnick

J. H.

Fyfe

J. J.

. 1993. Above the Law: Police and the Excessive Use of Force. New York: Free Press.

73.

StataCorp. 2009a. Stata Statistical Software: Release 11. College Station, TX: StataCorp LP.

74.

StataCorp. 2009b. Stata 11 Multiple-imputation Reference Manual. College Station, TX: Stata Press.

75.

StataCorp. 2011. Stata Multiple-imputation Reference Manual: Release 12. College Station, TX: StataCorp LP.

76.

Stevens

Cho

J. H.

. 1985. “Socioeconomic Indexes and the New 1980 Census Occupational Classification Scheme.” Social Science Research 14(2):142–168.

77.

Sweeten

2012. “Scaling Criminal Offending.” Journal of Quantitative Criminology 28:533–57.

78.

Terrill

. 2003. “Police Use of Force and Suspect Resistance: The Micro Process of the Police-suspect Encounter.” Police Quarterly 6:51–83.

79.

Terrill

Mastrofski

S. D.

. 2002. “Situational and Officer-based Determinants of Police Coercion.” Justice quarterly 19:215–48.

80.

Thaler

R. H.

Tversky

Kahneman

Schwartz

. 1997. “The Effect of Myopia and Loss Aversion on Risk Taking: An Experimental Test.” The Quarterly Journal of Economics 112:647–61.

81.

Torrey

E. F.

Kennard

A. D.

Eslinger

Lamb

Pavle

. 2010. More Mentally Ill Persons Are in Jails and Prisons than Hospitals: A Survey of the States. Arlington, VA: Treatment Advocacy Center.

82.

Tversky

Kahneman

. 1992. “Advances in Prospect Theory: Cumulative Representation of Uncertainty.” Journal of Risk and Uncertainty 5:297–323.

83.

Tyler

T. R.

1990. Why People Obey the Law. New Haven, CT: Yale University Press.

84.

Tyler

T. R.

1997. “The Psychology of Legitimacy: A Relational Perspective on Voluntary Deference to Authorities.” Personality and Social Psychology Review 1:323–45.

85.

Tyler

T. R.

2011. “Trust and Legitimacy: Policing in the USA and Europe.” European Journal of Criminology 8:254–66.

86.

Tyler

T. R.

Blader

S. L.

. 2000. Cooperation in Groups: Procedural Justice, Social Identity, and Behavioral Engagement. Philadelphia, PA: Psychology Press.

87.

U.S. Department of Justice, Bureau of Justice Statistics, Federal Bureau of Prisons. 2007. Survey of Inmates in State and Federal Correction Facilities, 2004 [Computer file]. Compiled by U.S. Department of Commerce, Bureau of the Census. Ann Arbor, MI: Inter-Consortium for Political and Social Research.

88.

U.S. Department of Justice, Federal Bureau of Investigation. 2009. Uniform Crime Reporting Program Data: Arrests by Age, Sex, and Race, 2009. ICPSR30761-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2011-09-30. doi: 10.3886/ICPSR30761.

89.

U.S. Department of Justice. Federal Bureau of Investigation. Criminal Justice Information Services Division. 2012. “Uniform Crime Report: Law Enforcement Officers Killed and Assaulted, 2011.” Retrieved April 5, 2013 (http://www.fbi.gov/about-us/cjis/ucr/leoka/2011/officers-assaulted-1/assaults_2011.pdf).

90.

U.S. Department of Justice, Federal Bureau of Investigation, Criminal Justice Information Services Division. 2014. “Uniform Crime Report: Law Enforcement Officers Killed and Assaulted, 2013.” Retrieved September 17, 2015 (https://www.fbi.gov/about-us/cjis/ucr/leoka/2013/officers-assaulted/assaults_topic_page_2013.pdf).

91.

Van Maanen

1978. “The Asshole.” Pp. 221–38 in Policing: A View from the Street, edited by Manning

P. K.

Van Maanen

. Chicago, IL: Goodyear.

92.

Wallace

Mullen

P. E.

Burgess

. 2004. “Criminal Offending in Schizophrenia over a 25-year Period Marked by Deinstitutionalization and Increasing Prevalence of Comorbid Substance Use Disorders.” American Journal of Psychiatry 161:716–27.

93.

Watson

A. C.

Angell

. 2013. “The Role of Stigma and Uncertainty in Moderating the Effect of Procedural Justice on Cooperation and Resistance in Police Encounters with Persons with Mental Illnesses.” Psychology, Public Policy, and Law 19:30.

94.

Weidner

Robert R.

Terrill

William

. 2005. “A Test of Turk’s Theory of Norm Resistance Using Observational Data on Police-suspect Encounters.” Journal of Research in Crime and Delinquency 42:84–109.

95.

Worden

R. E.

Shepard

R. L.

. 1996. “Demeanor, Crime, and Police Behavior: A Reexamination of the Police Services Study Data*.” Criminology 34:83–105.