An Exploratory Analysis of Time Lapses in Serving Arrest Warrants

Abstract

The seizure of persons through the authority of an arrest warrant issued by a magistrate is the only method of arrest specifically enumerated in the U.S. Constitution, yet scant research exists on the processing and serving of arrest warrants. This study investigated the correlates of how long it took a sample of arrest warrants to be served by law enforcement officials. Using focal concerns and disproportionate police contact underpinnings, the findings revealed the characteristics predictive of rapid service of warrants in this jurisdiction were weakly associated with the focal concerns perspective and generally unrelated to the disproportionate contact perspective. The findings highlight the need for more detailed study of, and the development of, better theoretical foundations for this issue.

Keywords

arrest warrant sheriff police focal concerns disproportionate contact

The only method of arrest described in the U.S. Constitution, enumerated in the Fourth Amendment, is an arrest with a warrant, based upon probable cause, and supported by an oath or affirmation. It is therefore interesting that scant empirical research exists on the issuing, processing, and serving of arrest warrants in the United States. This dearth of research is regretful and has even attracted the attention of the U.S. Senate, which held hearings in 2001 on the national backlog of arrest warrants (U.S. Senate, Subcommittee on Criminal Justice Oversight, 2001). At that time, the Subcommittee on Criminal Justice Oversight issued a statement that the nation faces an extreme backlog of unserved arrest warrants, with fugitives on the loose, continuing to commit new crimes (U.S. Senate, Subcommittee on Criminal Justice Oversight, 2001, p. 1).

It was estimated that the unserved arrest warrants in the United States at any point in time exceeded 3 million in the early 2000s (Officer of the Inspector General of Audit, 2000; U.S. General Accounting Office, 2001). The enormity of unserved arrest warrants poses serious problems for the justice system and society, as the courts cannot operate efficiently or effectively if criminal defendants cannot be made to appear. When the application of sanctions by the court is indefinitely delayed, the criminal justice process lacks any hope of deterrence. Moreover, there is no justice for victims when defendants avoid court, and because some arrest warrants stem from probation and parole violations, the warrant backlog interferes with the efficient operation of correctional agencies.

This reveals a need to investigate why some fugitives are apprehended quickly while others are not and what case characteristics are associated with how quickly a warrant is served. As a matter of justice, it is important to examine whether fugitive demographic characteristics, such as age, sex, and race, explain the rapidity of serving arrest warrants. It would be more appropriate if practical public safety concerns, such as seriousness of offense and the seriousness of the fugitives past criminal history, explained the rapidity with which arrest warrants were served. The present study involved an initial examination of the correlates to the time taken to serve arrest warrants.

Borrowing from the conflict and focal concerns theoretical perspectives, the present study examines correlates of the length of time taken to serve a sample of 428 arrest warrants issued by one county court. Relying on a focal concerns perspective, case characteristics associated with blameworthiness, public safety, and practical constraints were examined as predictors of length of time taken to serve the warrant. Exercising a disproportionate police contact perspective, fugitive characteristics associated with age, race, and gender were also explored as predictors. Correlates were examined to determine whether one or both of these perspectives helped explain how quickly warrants were served.

Literature Review

Our empirical knowledge about arrest warrants is extremely limited. Scant research exists on how arrest warrants are issued, processed, or served. This fact was lamented in 2001 by the Subcommittee on Criminal Justice Oversight of the U.S. Senate, which called for further social scientific research into the prevalence and processing of arrest warrants. Unfortunately, not much research has occurred in this area since this call was made. In May 2013, literature searches were conducted through the Social Science Index, Criminal Justice Abstracts, Google Scholar, and National Criminal Justice Reference Service Abstracts databases to locate any social scientific studies regarding the issuance, processing, and serving of arrest warrants. Several articles from police trade journals were located that involved anecdotal discussions of fugitive apprehension units, but only a small handful of empirical works were located. Most of these empirical works were published as governmental reports, simply presenting descriptive statistics on the prevalence and nature of arrest warrants, and failed to include any multivariate analyses. These descriptive studies suggest that the number of arrest warrants issued annually in the nation, and the backlog of unserved warrants, is enormous.

Arrest Warrant Prevalence

Although it is difficult to know the exact number of outstanding arrest warrants that exist across the United States (Flannery & Kretschmar, 2012), estimates suggest that the number is large and growing. For example, a report by the U.S. General Accounting Office (2001) indicated that in August 2001, there were over 825,000 active outstanding arrest warrants in the National Crime Information Center (NCIC) database. While a very large number, it is only a fraction of the unserved arrest warrants across the nation. Being a federal database, the warrants in NCIC are primarily from federal cases, and the most serious cases are from state and local courts. For example, another audit found that only 28% of the arrest warrants found in state level, crime computer databases were also listed in the NCIC database (U.S. Office of the Inspector General of Audit, 2000). If this is consistently the case, it suggests that there may be nearly 3 million unserved arrest warrants in the United States at any point in time.

Available research also suggests the backlog of unserved warrants continues to grow. In a study funded by the National Institute of Justice, Guynes and Wolff (2004) reviewed arrest warrant records in Montgomery County, Maryland, and Hennepin County, Minnesota, and found that each county experienced a net increase in the number of unserved warrants over the 6-month study period. Specifically, at the start of their analyses, Montgomery County had 11,646 unserved arrest warrants in its computer system, and Hennepin County had 25,411. During the 6-month period of study, 6,428 warrants were served in Montgomery County, but an additional 6,544 new warrants were issued, increasing the backlog of unserved warrants by 116. Hennepin County started the study with 25,411 unserved warrants and increased its backlog by 917 warrants after 6 months. This is consistent with Hager, Daniel, Graycarek, and Knowles’s (2005) statewide evaluation of unserved arrest warrants in Kentucky. At the start of the study, they estimated Kentucky had as many as 265,000 unserved arrest warrants, and over the 3-year period of study, 568,606 arrest warrants were issued within that state. While many arrest warrants were served by law enforcement officials annually, the state’s backlog of unserved warrants grew by approximately 28,000 warrants per year. A similar analysis conducted by White (2000) revealed that in October 1999, the courts in Milwaukee County, WI, had 49,366 unserved warrants in its computer database, and another 557 warrants were issued in the next month. While it is clear that the backlog of unserved arrest warrants in the United States is extensive and growing, it is important to consider the source of the backlog.

Arrest Warrant Characteristics

The primary contributor to the growing backlog appears to be warrants issued for failing to appear in court rather than for the original charge (which is referred to as an “instanter” charge). Using court case data for an 11-year period from 1994 through 2005, Goldkamp and Vilcica (2008) found that the number of unserved failure to appear warrants in that city fluctuated between 40,000 and 50,000, and on average, 20,900 new warrants for failing to appear were issued annually in Philadelphia, PA. In order to add perspective to these numbers, they revealed that these courts only received an average of about 35,000 new criminal cases annually. Thus, more than half of new cases ended up having a warrant issued for the defendant’s failure to appear in Philadelphia, which is consistent data from other jurisdictions (Guynes & Wolff, 2004; Hager, Daniel, Graycarek, & Knowles, 2005; White, 2000).

Only about a quarter of unserved warrants appear to be issued for an instanter charge (Guynes & Wolff, 2004; Hager et al., 2005; White, 2000), and, of these, very few originate from cases involving violent felonies (Guynes & Wolff, 2004; White, 2000). Aside from failure to appear and new charges, unserved warrants might also originate from traffic or civil cases (Guynes & Wolff, 2004; White, 2000), violating conditions of probation or parole (Guynes & Wolff, 2004; White, 2000) or other noncriminal violations such as local ordinance violations (Hager et al., 2005). Regardless of the source of the backlog, the number of arrest warrants in the United States that remained unserved at any point in time is enormous, and the backlog appears to be expanding.

Remaining Gaps in the Research

The literature pertaining to arrest warrants is limited and constrains the field’s understanding regarding their issuance, processing, and serving. Most notably, only a few studies have examined these important decisions relating to arrest warrants. Moreover, the studies that do exist are mostly published by government entities (e.g., Kentucky Legislative Research Commission) or private agencies (e.g., Wisconsin Policy Research Institute) and are descriptive in nature. While these reports are interesting and provide general information about arrest warrants, they fail to rigorously examine factors influencing their issuance, processing, or serving in a meaningful way. Based on the theoretical underpinnings from two different frameworks, the current study uses a multivariate analysis to explore whether certain factors, both legal and extralegal, influence the timeliness with which arrest warrants are served.

Theoretical Frameworks

Unlike the study of offender behavior, much criminal justice system research has failed to incorporate a theoretical foundation for the gathering of data and the interpretation of results (Bernard & Engel, 2001; Duffee & Maguire, 2007; Kraska & Brent, 2011). Explicitly stated theoretical frameworks are vital for gaining a better understanding of criminal justice agents’ and agency behaviors (Engel, Calnon, & Bernard, 2002). One exception, however, has been the literature on criminal sentencing (Johnson, Ulmer, & Kramer, 2008; Steffensmeier, Kramer, & Streifel, 1993; Steffensmeier, Ulmer, & Kramer, 1998), which has utilized a focal concerns theoretical perspective to guide the selection of variables and interpretation of results.

Focal Concerns

Originally, focal concerns theory was employed to explain sentencing decisions of judges (Steffensmeier, 1980), but, more recently, scholars have applied the framework to decisions made in policing in an effort to contextualize use of force (Crow & Adrion, 2011) and traffic stop (Higgins, Vito, & Grossi, 2012) outcomes.¹ In its judicial application, the theory posits that judges base their sentencing decisions on three focal concerns—blameworthiness, protection of the community, and practical constraints and consequences (Steffensmeier, 1980; Steffensmeier, Ulmer, et al., 1998). Blameworthiness refers to an individual’s culpability and having the punishment fit the crime. Protection of the community relates to incapacitation of the defendant based on the individual’s risk of future dangerousness or their perceived risk of recidivating. Concerns centered on practical constraints and consequences relate to the organizational costs, legal restrictions, and limited space experienced by the criminal justice agencies involved. Among these three main focal concerns, protection of the community is often the most important in determining sentences (Steffensmeier, Ulmer, et al., 1998).

As applied to the judiciary, focal concerns theory takes into account the difficulty in making decisions in cases in an environment marked by ambiguity, uncertainty, and massive caseloads (Albonetti, 1991; Steffensmeier, 1980). Uncertainty stems from conflicting sentencing goals, a lack of information about the defendant(s), and the difficulty predicting the risks of recidivism. Because judges are often forced to make decisions with incomplete information on the true nature of the defendant and case (Albonetti, 1991) they may resort to “perceptual shorthand” to guide their decisions. These predesigned decision-making schemes are influenced by certain legal (e.g., seriousness of the offense, prior record, weapon use, etc.) and extralegal (e.g., race, gender, age, etc.) factors pertaining to the defendant and his/her case (Johnson, 2003; Steffensmeier, Ulmer, et al., 1998). The suggestion that decision makers rely on extralegal factors is not new to policing research. Jerome Skolnick’s (1994, p. 12) suggestion that people need not actually be criminal, but “merely conform to the stereotype,” to be adjudged criminal reflects the idea that criminal involvement is associated with certain individual characteristics that are not based on legal relevance. Similarly, Smith and Alpert (2007) contend that officer behavior is sometimes motivated by unconscious stereotypes that develop as a result of repeatedly encountering similarly situated people in similar contexts.

Research indicates that legal factors are the most influential in determining judicial decision making (Steffensmeier, Ulmer, et al., 1998), but, importantly, extralegal factors also contribute to judges’ decisions. Thus, net controls (i.e., legal factors), extralegal factors influence decisions, resulting in certain groups experiencing differential treatment, despite their cases being similar from a legal perspective (i.e., offense seriousness, prior history, etc.).² This suggests that while the bulk of consideration is given to legally relevant factors, extralegal factors, such as the defendant’s race/ethnicity, age, gender, and social class, also influence judges’ decisions and explain differences in outcomes (DeMuth, 2003; DeMuth & Steffensmeier, 2004; Johnson, 2003; Steffensmeier, Ulmer, et al., 1998; Ulmer & Johnson, 2004).

A focal concerns approach can also be applied to explain the decisions of law enforcement personnel to focus their limited resources on trying to serve specific arrest warrants. Law enforcement agencies primarily tasked with serving arrest warrants have limited resources and the serving of arrest warrants must be balanced against all of the other duties of the agency (Hager et al., 2005). Only a limited number of personnel can be assigned to the task of serving arrest warrants, and these personnel may also have other responsibilities. Finally, this situation is further confounded by the limited background information warrants sometimes offer on the life and behavior of the fugitive. These elements can be compared to the environment of the sentencing judge in terms of case ambiguity, uncertainty, and workload (Albonetti, 1991; Steffensmeier, 1980).

It can be presumed that law enforcement personnel deal with this environment by prioritizing the service of some warrants over others based on the three focal concerns. Protection of the community, in this context, would relate to the fugitive’s perceived risk of future dangerousness while remaining at large. Having limited information on hand about the defendant, law enforcement officers may rely on certain case characteristics to determine the risk of future dangerousness. They may also rely on characteristics of the defendant’s past criminal history as proxy for causing future harm to society. Because protection of the community is paramount, arrest warrants for crimes against persons, felony cases, and criminal cases may be served more quickly than warrants from traffic and civil cases, misdemeanors, and nonpersonal offenses. Likewise, warrants for persons with lengthy criminal records of prior offenses may be served more rapidly than warrants for fugitives with few prior arrests. It is important to note that one’s “criminal history can be used as an indicator of both the blameworthiness and community safety concepts” (Hartley, Madden, & Spohn, 2007, p. 63).

The focal concern of blameworthiness refers to the fugitive’s culpability and may also be estimated through case characteristics. Warrants for failures to appear may suggest to officers that the court does not perceive the fugitive as highly culpable or dangerous, since the accused was allowed to return to the community before trial, rather than remaining in preventive detention in jail. Instanter warrants for an original charge, on the other hand, may signal to officers that the defendant is an active offender deserving of immediate attention. Another characteristic that may be viewed this way is the amount of bond attached to the warrant. The amount of bond required by the judge may indicate to the officer how culpable the judge perceived the fugitive. Additionally, the defendant’s criminal history serves as a proxy for culpability (DeMuth, 2003; Hartley et al., 2007; Sevigney & Caulkins, 2004), in that repeat offenders are usually viewed as being more blameworthy for their behavior because they have previously been punished (von Hirsch, 1976). In fact, federal sentencing guidelines explicitly state, “A defendant with a record of prior criminal behavior is more culpable than a first offender and thus deserving of greater punishment” (U.S. Sentencing Commission, 2013). As a result, instanter warrants, warrants with high bond amounts, and warrants issued for repeat offenders may be served more quickly than warrants from other later stages of the case, with lower bond amounts, or nonfrequent offenders.

Officers may encounter practical constraints and consequences in several forms when trying to serve warrants. These may include the number of personnel available to devote to serving arrest warrants or the information available on the identity and whereabouts of the fugitive. Whether or not the fugitive is outside the jurisdiction is another practical constraint due to problems of out-of-state extradition and in-state transportation. Arrest warrants may be served quickly when the defendant is residing within the issuing court’s jurisdiction, but considerable time might elapse between issuance and serving for defendants who live out of jurisdiction.

Disproportionate Police Contact

In some instances, officers are not actively seeking to serve an arrest warrant but through the course of their daily activities accidentally encounter a fugitive (Hager et al., 2005; White, 2000). In scenarios such as this, officers generally do not know of the existence of the warrant before they contact the fugitive and, therefore, the focal concerns perspective does not apply to warrants served in this manner. Consequently, an alternative theoretical framework should be utilized to explain how quickly warrants are served when their discovery is accidental. A conflict perspective based on disproportionate police contact helps explain why some fugitives are more likely to have contact with officers than others.

Jerome Skolnick (1966) suggested that because of the ever-present possibility of being attacked, police officers become suspicious of the routine aspects of daily life, alert to signs that something might not be right. Police officers, made distrustful by experience, develop a sense of caution and suspicion that produces the notion of the “symbolic assailant.” Officers develop perceptions of certain kinds of people as symbolic assailants, that is, as persons whose appearance and behavior officers have come to associate with violence or criminal behavior (Skolnick, 1966). This leads persons who display the characteristics of the symbolic assailant to be at risk for disproportionate scrutiny by the police, including being stopped and checked for warrants. It is likely that wanted individuals who experience disproportionate contact with the police because they fit the profile of a symbolic assailant, therefore, will have their warrants served more quickly than individuals who lack characteristics that may bring heightened police scrutiny.

Recent research on police–public contacts revealed several characteristics that correlate with elevated levels of contact with the police. A national survey of 76,910 drivers by the Bureau of Justice Statistics in 2002 revealed that 8.7% had been stopped by the police within the last year (Smith & Durose, 2006) and those who experienced a stop by the police differed in many ways from those who had not been stopped. Those stopped were almost twice as likely to also report experiencing other types of face-to-face contact with the police, confirming that many of these persons had disproportionate police contact. Drivers stopped were more likely to be male and under age 25, but driver race was not related to likelihood of being stopped (Smith & Durose, 2006).

Pollock, Oliver, and Menard (2012) reviewed 16 quantitative studies and 7 qualitative studies on police–public contact. These studies produced many contradictory findings about whom experienced disproportionate contact with the police, but many of the studies indicated that non-Whites had more contact than did Whites, males had more contact than did females, and juveniles and young adults had more contact than did older adults, and those with lengthy prior criminal histories had more contact than those with shorter records. Pollock and Associates (2012) then proceeded to analyze survey data from more than 13,000 incidences of police contact. Their results also revealed that disproportionate contact with the police was best explained by being male and having a delinquent past.

These findings suggest that the wanted individuals most likely to have contact with police officers in any circumstance are those who are young, male, and with lengthy criminal histories. To a lesser extent, wanted individuals who are African American or Hispanic/Latino may have more police contact (Anderson, 1990; Chambliss, 1994; Harris, 1997). Arrest warrants, therefore, may be served quickest for individuals who have lengthy criminal records and are male, young, and persons of color.

Hypotheses

These perspectives suggest two general research hypotheses. First, based on a focal concerns perspective, it is predicted that arrest warrants will generally be served more quickly when the characteristics of the warrant and the defendant suggest a risk to public safety (criminal case, felony case, crime against a person, and lengthy criminal history), blameworthiness (instanter warrant, bond amount, and lengthy criminal history), and lack of practical constraints (defendant resides within the jurisdiction).

Hypothesis 1: Warrants will be served more quickly when the risks to public safety, level of blameworthiness, and lack of practical constraints are highest, net controls.

The second perspective, warrants served as a by-product of disproportionate scrutiny by the police, suggests a second research hypothesis. It is expected that arrest warrants will generally be served more quickly on defendants who characterize persons who experience disproportionate contact with the police, namely, younger, non-White males.

Hypothesis 2: Warrants will be served more quickly when the defendant most resembles the “symbolic assailant,” net controls.

These general research hypotheses were examined using a sample of arrest warrants served in one county in the United States.

Method

Data were gathered from one county court located in a Midwestern state. The county had a population of more than 300,000 residents and was part of a metropolitan area with a population in excess of 2 million persons. The racial composition of the county was 85% Caucasian, 2% African American, 8% Hispanic/Latino, and 5% other racial groups. The Hispanic/Latino population was very transitory, being largely composed of immigrants employed in the locally booming housing industry and in the seasonal agricultural industry. The median household income, median home value, and average education level were above the state average. The primary industries in the county were agriculture, recreation, and manufacturing.

Within this county, all arrest warrants issued by the county courts were issued to, and held by, the sheriff department. The sheriff department only allowed us access to warrant files that had been closed and were awaiting destruction by a document shredding company. We were not granted to currently active warrants because, being the original court document, they needed to be accessible to law enforcement officers and court personnel at all times. The closed warrant files numbered in the thousands so, due to time constraints, we chose to select a random sample of 1,000 closed files for arrest warrants issued between December 1, 2009, and June 30, 2010. This was conducted with SPSS (version 21) by assigning each case a random number and then querying SPSS to select a sample of 1,000 cases. The sheriff department personnel then retrieved these selected cases for our examination.

Unfortunately, and to our surprise, a large number of the warrant files we selected lack sufficient information for our analysis. Warrants that had been closed through some other means than arrest, such as recall by the court, or the surrender or death of the wanted individual, made up 57.2% of the original 1,000 cases. Files for the 572 cases that had not been closed by an arrest contained only the warrant document and a brief, handwritten court order, indicating that the warrant had been rescinded. Neither of these documents consistently offered much more descriptive information on the defendant than the defendant’s name, date of birth, and social security number, preventing us from collecting useful quantitative data about the case. The bulk of these warrants (540 cases, 94.4%) were quashed by the court for a number of legal reasons, such as the case being too old due to a statute of limitation, the prosecutor deciding to drop charges, or the plaintiff (in civil cases) withdrawing the case. As best we could decipher, based on the handwritten court documents, only 5.6% of the files (32 cases) were closed due to the defendant’s voluntary surrender or death.

On the other hand, 428 of the warrants from the original 1,000 cases had been served by a law enforcement officer. These case files contained the warrant document, a booking sheet from the jail, and a state criminal history computer printout for the defendant, dated with the time of the apprehension. These documents provided sufficient details about the defendant, and his or her apprehension, to facilitate the creation of quantitative variables for our study. Therefore, our final sample included all 428 arrest warrant cases that had been served by law enforcement officers.

Measures

The dependent variable of study was the length of time taken to serve each arrest warrant. This was measured by the number of days between the day the warrant was issued by the court and the day the warrant was served. Shorter warrant service times were assumed to be indicative of defendants prioritized by officers for warrant service or targeted by officers for further scrutiny on the street.

The selection of the exogenous variables was guided by the focal concerns and disproportionate police contact underpinnings. As previously described above, we predicted that, due to the focal concern of danger to public safety, arrest warrants for crimes against persons, felonies, and criminal cases would be served more quickly than warrants from traffic and civil cases, misdemeanors, and nonpersonal offenses. Thus, three dichotomous variables were created. In the first, warrants for crimes against a person were coded 1, and all other offenses (such as property crimes, drug offenses, public order offenses, and civil offenses) were coded 0. In the second measure, warrants for felony offenses were coded 1, and misdemeanors and civil offenses were coded 0. In the third measure, criminal cases were coded 1, and civil offense warrants (such as civil body attachment warrants) were coded 0.

The second focal concern was blameworthiness, or culpability of the defendant (Steffensmeier, Kramer, et al., 1993, Steffensmeier, Ulmer, et al., 1998; Ulmer & Johnson, 2004), and it was predicted that instanter warrants, those with high bond amounts, and those for fugitives with lengthier criminal records would be served more quickly than warrants from other stages of the case, those with lower bond amounts, and those for individuals with fewer prior offenses. Three variables were used to measure these characteristics. The first measured whether the warrant was an instanter warrant, as opposed to a warrant associated with the continuation of an existing case, such as a warrant for failing to appear for a court date or a violation of a condition of one’s sentence.³ This was measured dichotomously. The second measure was the dollar amount of the bond assigned to the arrest warrant. The third measure was the total number of prior criminal arrests on the wanted individual’s criminal record.

The third focal concern was practical constraints, which was based on the potential limitations experienced by law enforcement personnel when trying to track down and extradite defendants outside of their jurisdiction. Information was available to determine whether the wanted individual was apprehended within the county that issued the arrest warrant. This measure of residing within the jurisdiction was coded 1 if the wanted individual was apprehended within the county and 0 if apprehended outside the county.⁴ Unfortunately, this was the only measure of practical constraints utilized in the present study, as no data were available on such administrative conditions, for example, the number of officers available at any given time to serve warrants, other officer workload responsibilities, and police financial resources.

Exogenous variables were also created to measure the likelihood of disproportionate police contact, although some theoretical constructs did overlap with the measures of focal concerns. For example, prior criminal record is considered a predictor of disproportionate police contact as well as culpability and risk to public safety measures of focal concerns (DeMuth, 2003; Hartley et al., 2007; Sevigny & Caulkins, 2004).

Males commit proportionately more crime and make up the vast majority of all arrests and drivers stopped by the police, suggesting males would be apprehended faster than females. Warrants for male individuals were coded 1 and females 0. While the evidence is equivocal (Pollock, 2012; Pollock, Oliver, & Menard, 2012; Smith & Durose, 2006), many scholars have argued persons of color, especially African Americans and Hispanics, are often stereotyped as criminals by the police, resulting in disproportionate numbers of stops, searches, and arrests among these peoples (Anderson, 1990; Chambliss, 1994; Harris, 1997). The racial classifications listed on the wanted individuals’ criminal histories were used to create two dummy variables for the racial groups Black and Hispanic. All other non-White racial groups made up 0.9% of the total sample, so these cases were merged with Whites as the reference category.⁵ Finally, the wanted individuals’ ages were calculated as their age on the date the arrest warrant was issued, with the expectation that younger adults would draw more scrutiny by the police than older adults.

Analysis

As the first step of the analysis, descriptive statistics were calculated and examined. Next, bivariate analyses and tests for multicollinearity were conducted among the independent variables. Finally, multivariate analyses were conducted using Cox survival analysis models to test the predictions discussed above, while simultaneously controlling the net effects of the other variables in the model. Cox proportional hazards models are semiparametric and do not require any assumptions regarding the shape of the dependent variable’s distribution, yet still utilize continuous measures rather than categorical measures (Allison, 1984; Walters, 2009). It estimates a linear-like model and allows estimates of covariates without specifying the precise form of the dependent variable, other than a measure of time before the occurrence of a specific event (Therneau & Grambsch, 2000).

The Cox proportional hazards test models the time taken for events to occur, often called “survival analysis” (Cox, 1972; Therneau & Grambsch, 2000). Survival models relate to the time that passes before some event occurs (such as a warrant being served) and the relationships between the event and covariates (the independent variables) that may be associated with when that event occurs. Often used in medical research regarding terminal illnesses, a Cox proportional hazards model provides an estimate of a treatment effect on patient survival after adjusting for other explanatory variables. It allows the estimation of the risk (or hazard) of death for a patient, given their prognostic variables (Walters, 2009, p. 1). Applied here, it allowed the estimation of the likelihood of an arrest warrant being served within a certain window of time, given the specific characteristics of the warrant case.

Results

Table 1 displays the minimum and maximum values, means, and standard deviations for each of the variables used in the study. The dependent variable, the number of days taken to serve the arrest warrant, ranged from 0 days to approximately 9 years. This produced a skewed distribution with a mean of approximately 143 days, but a median of 20 days. The wanted individuals were overwhelmingly male, a quarter Black or Hispanic, with a mean age in the early 30s. Most had many prior criminal arrests (seven on average), and the majority of the warrants stemmed from failing to appear for court or violating sentencing conditions on a nonviolent misdemeanor or civil case. Examination of the descriptive statistics in Table 1 revealed that the number of prior arrests and the bond amount variables had skewed distributions. These two variables, therefore, were transformed to their natural logs for the remaining analyses.

Table 1.

Sample Descriptive Statistics and Tests of Mean Differences.

Variable	Min–Max	Mean (SD)
Male	0–1	.77 (.42)
Black	0–1	.07 (.26)
Hispanic	0–1	.18 (.39)
Age	18–59	32.46 (10.06)
Logged total prior criminal arrests	0–3.83	1.45 (1.11)
Instanter case	0–1	.28 (.45)
Criminal case	0–1	.23 (.42)
Felony case	0–1	.23 (.42)
Offense against persons case	0–1	.11 (.32)
Logged bond amount	4.61–11.92	8.12 (1.48)
Defendant apprehended out of county	0–1	.39 (.49)
Dependent variable
Number of days taken to serve warrant	0–3,289	142.90 (345.91)

Note. N = 428.

Significance levels: *p < .05. **p < .01. ***p < .001.

Bivariate Analyses

Pearson’s r correlations were explored, examining for statistically significant bivariate relationships with the endogenous variable and for collinearity issues among the exogenous variables. Table 2 reveals the results of these correlation tests. At the bivariate level, 5 of the 11 exogenous variables reached statistical significance in their correlations with the number of days taken to serve the arrest warrant. Hispanic defendants generally took longer to be apprehended than defendants of other races. Defendants with lengthier criminal histories were apprehended faster than defendants with fewer prior arrests. Warrants for instanter charges were served faster than warrants stemming from later in the case’s history due to failures to appear or violations of one’s sentence conditions. Warrants with higher bonds took longer to serve than warrants with lower bonds. Warrants for persons apprehended living outside of the county took longer to serve than warrants for persons apprehended within the county that issued the warrant. The seven remaining exogenous variables failed to reach statistical significance at the p < .05 level, at least in a bivariate analysis before the net effects of the other variables were controlled.

Table 2.

Bivariate Correlation Matrix.

	1	2	3	4	5	6	7	8	9	10	11	12
1. Days to serve	1.00
2. Male	0.05	1.00
3. Black	0.02	0.03	1.00
4. Hispanic	0.20**	0.03	−0.13**	1.00
5. Age	0.06	0.05	0.01	−0.07	1.00
6. Logged arrests	−0.11*	0.12*	0.04	−0.05	0.05	1.00
7. Instanter	−0.13**	−0.03	0.02	−0.02	0.08	0.15**	1.00
8. Criminal case	−0.08	−0.04	0.04	−0.12*	−0.12*	0.19**	0.35**	1.00
9. Felony case	0.04	−0.03	0.18**	−0.15**	−0.02	0.09	0.33**	0.47**	1.00
10. Personal crime	−0.09	0.09	−0.05	−0.01	−0.02	0.29**	0.18**	0.32**	0.08	1.00
11. Logged bond	0.19**	−0.01	0.16**	−0.03	−0.14*	0.11*	−0.01	0.21**	0.37**	0.06	1.00
12. Out of county	0.24**	0.11*	0.24**	0.14**	−0.08	0.11*	−0.09	0.20**	0.08	−0.16**	0.07	1.00

Note. N = 428.

Significance levels: *p < .05. **p < .01.

The correlation matrix also suggested that there were no issues with collinearity among the exogenous variables, as all of these bivariate relationships produced Pearson’s r coefficients of less than .50. This was further confirmed by the examination of their variance inflation factors (VIFs) in the multivariate model. The multivariate VIF values of the 11 exogenous variables were each below the conventional threshold of 5, thus confirming that collinearity among the independent variables was not an issue (Allison, 1999, p. 142).

Multivariate Analyses

The multivariate analysis technique was driven by the nature and distribution of the endogenous variable. First, the dependent variable’s distribution was highly skewed (skewness coefficient = 5.106, Shapiro–Wilk’s normality test: p > .05; Shapiro & Wilk, 1965). The survival curve distribution of the dependent variable for the entire sample is displayed in Figure 1. While the median of this distribution for the entire sample was 20 days, it took 112 days before 75% of the warrants were served, and the longest time to serve a warrant was 3,289 days. This dramatically skewed distribution violated one of the assumptions necessary for estimating linear regression models, prohibiting the use of ordinary least squares regression (Miles & Shevlin, 2001). Because the dependent variable was also a continuous level of measurement, rather than categorical, neither was the use of logistic regression models the best option (Cameron & Trivedi, 1998). Finally, the dependent variable was essentially a measure of case survival (i.e., how long it took each warrant case to be cleared). The best multivariate technique for use here, therefore, was the Cox proportional hazards regression model (Cox, 1972).

Figure 1.

Survival curve of the dependent variable (time to warrant being served).

A Cox model was estimated for the time taken to serve an arrest warrant using the exogenous variables as covariates. The regression coefficients, standard errors, hazard ratios, and statistical significance levels of these covariates are displayed in the model in Table 3. When examining the unstandardized regression coefficient for each of the exogenous variables, a negative coefficient indicates that the exogenous variable reduces the time taken to serve the warrant. Conversely, positive coefficients imply that the exogenous variable serves to lengthen the amount of time taken to serve the warrant. Finally, the hazard ratio represents a standardized change in the likelihood of the warrant being served, after adjusting for the other explanatory variables in the model (Therneau & Grambsch, 2000; Walters, 2009).

Table 3.

Cox Proportional Hazards Multivariate Models.

Predictor	Coefficient (SE)	Hazard ratio	Significance
Male	.256 (.124)	1.292	.039
Black	−.077 (.205)	0.926	.707
Hispanic	.518 (.136)	1.679	.000
Age	−.010 (.005)	0.990	.056
Logged prior arrests	.081 (.047)	1.085	.085
Instanter warrant	−.434 (.142)	0.648	.002
Criminal case	−.026 (.144)	0.974	.857
Felony case	−.300 (.173)	0.741	.023
Crime against persons case	−.188 (.170)	0.828	.268
Logged bond	−.155 (.042)	0.856	.000
Defendant apprehended outside of county	.717 (.116)	2.049	.000
−2 log likelihood	4225.085
Model χ²	129.047***		.000

Note. N = 428.

In the model in Table 3, six explanatory variables reached statistical significance at the p < .05 level. When compared to the bivariate test results, after controlling for the net effects of the other exogenous variables in the model, defendant sex and felony cases became statistically significant, and logged prior arrests lost its statistical significance. Warrants for male defendants took an average of 1.3 times longer to serve than warrants for female defendants. Warrants for Hispanic defendants took an average of 1.7 times longer to serve than warrants for the reference category. Instanter warrants were generally served 35% faster than warrants stemming from later in the court case’s history. On average, felony case warrants were served about 26% faster than warrants from misdemeanor cases. Defendants apprehended living outside the county issuing the warrant took an average of 2.05 times longer to apprehend than defendants apprehended within the county of original jurisdiction.

In this multivariate model, the relationship bond amount had on time to serve the warrant changed direction from the relationship it demonstrated in the earlier bivariate analysis. In the bivariate analysis, cases with higher logged bond amounts took longer to serve than warrants with lower bond amounts. In the multivariate Cox regression analysis, however, cases with higher bond amounts were served faster than warrants with lower bond amounts. This change in direction was likely due to the ability of the Cox regression model to control for the net effects of the other variables in the model. For example, in Table 2 bivariate correlation matrix, logged bond amount demonstrated correlations with Black defendants (r = .16), the age of the defendant (r = −.14), number of prior arrests (r = .11), criminal cases (r = .21), and felony cases (r = .37). After controlling for the net effects of these other case characteristics, higher bond amounts were actually associated with faster bond service times.

Just as important were the exogenous variables that failed to reach statistical significance in this model. Black race, age, logged number of prior arrests, criminal case status, and crimes against person cases all failed to produce a statistically significant effect on warrant service time. Age and logged number of prior arrests did approach statistical significance, however, and would have if a more liberal p < .10 threshold had been used.

Discussion

The purpose of the present study was to examine potential correlates of how quickly arrest warrants are served. The exogenous variables examined were those suggested by the focal concerns perspective and the conflict perspective of disproportionate police contact. The focal concerns perspective, it was reasoned, suggested that priority would be given to warrants with case characteristics indicative of threats to public safety (e.g., lengthy criminal histories, personal crimes, felonies, and criminal vs. traffic and civil cases), blameworthiness (e.g., lengthy criminal histories, higher bond amounts, and instanter warrants), and lack of practical constraints (those located within the jurisdiction). Partial evidence was found to support this prediction.

For example, among the seven independent variables that could be associated with the focal concerns perspective, only four reached statistical significance in their hypothesized directions in the multivariate analysis. Instanter warrants, felony cases, bond amounts, and apprehension outside of the county all predicted the length of time taken to serve arrest warrants. A fifth variable associated with focal concerns—logged prior arrests—also approached statistical significance. Criminal case status (as opposed to a traffic or civil case) and cases involving crimes against persons failed to predict length of time to serve the warrant.

Nevertheless, even less evidence was found to support a conflict perspective of disproportionate police contact. No evidence was found that Blacks were apprehended more quickly than Whites, and the evidence here suggested that all races were apprehended faster than Hispanics on average. Female defendants were generally apprehended faster than men. Finally, although age approached but failed to reach statistical significance, the direction of the coefficient revealed that older defendants were generally apprehended faster than younger adult offenders. No evidence was found here to support the disproportionate contact perspective.

There may be multiple reasons for our findings of partial support for focal concerns and no support for disproportionate police contact. First, both of these perspectives might have simultaneously been in effect, canceling each other out statistically. For example, in most cases, arrest warrants are issued by the courts to a specific law enforcement agency (Guynes & Wolff, 2004; Hager et al., 2005). These specific agencies are charged with the primary responsibility of serving these warrants, usually accomplished by handing warrants out to patrol officers at roll call and expecting them to serve the warrants during the course of their shifts (Baker, 1985; Fletcher, 1990). Other times these agencies may have specialized warrant service units with officers assigned to do nothing but serve arrest warrants (Baker, 1985).

Contrast this with officers who discover a person has an active warrant for his or her arrest through other means. Patrol officers frequently learn, through computer checks, that the traffic law violators they have stopped are wanted on warrants (Baker, 1985; Fletcher, 1990). It is possible that the elements of a focal concerns perspective best explain the length of time warrants were served by law enforcement officers specifically tasked with serving warrants, and other perspectives would better explain warrant service times for warrants served by accidental discovery of the warrant. Unfortunately, there was insufficient information in the present data to identify the circumstance under which each warrant was served.

Another possible explanation for the present model’s impotence may have been that serving warrants was influenced more by the behaviors of the defendants than those of law enforcement officials. Specifically, this may relate to the actions the defendant may have taken, or was able to take, to avoid capture. Some support for this may be revealed in the effects of sex and Hispanic race. As female offenders are far more likely to be custodial parents than male offenders (Belknap, 2001), it is likely that they have fewer opportunities than male offenders to move frequently and stay unconnected with the broader society. The mobility of the Hispanic population of this county, being largely composed of migrant workers in the agriculture and construction industries, may have permitted Hispanic fugitives to leave the jurisdiction more easily than would be the case for Whites.

The possibility of defendants altering their behavior to evade detection is certainly a limitation that deserves mentioning, yet the previous literature offers conflicting support for this perspective. First, Goffman’s (2009) research suggested offenders might change their daily routines when they knew they were wanted, but it also indicates police resorted to a variety of savvy techniques to track down these wanted persons. Moreover, she intimates, but does not explicitly state, that if the police want someone badly enough they usually find them sooner or later as her accounts suggest that when her informants were on the run, they were eventually caught.

Second, this concern should be tempered with the enormous amount of research on offenders that suggests many possess low self-control (Gottfredson & Hirschi, 1990), have a penchant for drugs and alcohol (Karberg & James, 2005; Mumola & Karberg, 2006; Welte, Barnes, Hoffman, Wieczorek, & Zhang, 2005), suffer from mental health problems (Guy, Platt, Zwerling, & Bullock, 1985; James & Glaze, 2006), and are opportunistic (Wright & Decker, 1997). It is unlikely that many were capable of devising complicated schemes that would allow them to evade police attention for long, especially if police were actively trying to serve the warrants. Finally, the possibility that some fugitives might not know they are wanted and, therefore, would not alter their behavior cannot be ruled out. Data from 22 cities participating in the U.S. Marshals Service’s Fugitive Safe Surrender program revealed 18% of individuals who surrendered to the police did not actually have a warrant issued for their arrest (Flannery & Kretschmar, 2012). This finding suggested people’s knowledge of their warrant status is not always accurate.

A third explanation for the weak results could be that the present sample was atypical, highlighting the need for further research, especially with larger samples and samples with more geographic diversity. The present study, after all, had limitations in this regard. First, the study involved only one county, limiting the ability to generalize the findings. Second, there was a lack of information on warrants that were never served. It is unknown how these, permanently unserved warrants, differed from those actually served. Third, the accuracy of the data provided cannot be fully guaranteed. Inaccuracies in official agency data always exist (Klinger & Bridges, 1997; Maxfield, Lewis, & Szoc, 1980), and the extent of these inaccuracies (and their potential influence of the study’s findings) can never be fully known.

Notwithstanding the aforementioned limitations, the findings of this study do shed some initial light on how arrest warrants are served, but the findings also revealed the need for further study. Better theoretical foundations are needed for explaining the issuance and service of arrest warrants, and these theoretical foundations need to be empirically tested. Little is known about the judicial decision-making processes regarding this issuance of arrest warrants. What factors are correlated with the decisions of judges to issue arrest warrants rather than issue subpoenas and grant continuances? In what situations are judges likely to recall arrest warrants that have been issued? What factors determine whether specialized fugitive task forces serve arrest warrants rather than patrol officers? What circumstances encourage patrol officers to make a check for warrants on an individual? When do patrol officers proactively attempt to serve arrest warrants? Questions such as these have yet to be answered and open the door for a future research agenda on this most basic (and constitutionally based) method for making arrests.

Footnotes

Declaration of Conflicting Interests

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

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Albonetti

C. A.

(1991). Integration of theories to explain judicial discretion. Social Problems, 38, 247–266.

Allison

P. D.

(1984). Event history analysis: Regression for longitudinal event data. Beverly Hills, CA: Sage.

Allison

P. D.

(1999). Multiple regression: A primer. Thousand Oaks, CA: Pine Forge Press.

Anderson

(1990). Streetwise: Race, class, and change in an urban community. Chicago, IL: University of Chicago Press.

Baker

(1985). Cops: Their lives in their own words. New York, NY: Simon and Schuster.

Belknap

(2001). The invisible woman: Gender, crime and justice. Belmont, CA: Wadsworth.

Bernard

T. J.

Engel

R. S.

(2001). Conceptualizing criminal justice theory. Justice Quarterly, 18, 1–30.

Cameron

A. C.

Trivedi

P. K.

(1998). Regression analysis of count data. Cambridge, England: Cambridge University Press.

Chambliss

W. J.

(1994). Policing the ghetto underclass: The politics of law and law enforcement. Social Problems, 41, 177–194.

10.

Cox

D. R.

(1972). Regression models and life tables. Journal of the Royal Statistical Society, 34, 187–220.

11.

Crow

Adrion

(2011). Focal concerns and police use of force: Examining the factors associated with taser use. Police Quarterly, 14, 366–387.

12.

DeMuth

(2003). Racial and ethnic differences in pretrial release decision and outcomes: A comparison of Hispanic, black, and white felony arrestees. Criminology, 41, 873–907.

13.

DeMuth

Steffensmeier

(2004). Ethnicity effects on sentence outcomes in large urban courts: Comparisons among white, black, and Hispanic defendants. Social Science Quarterly, 85, 994–1011.

14.

Duffee

D. E.

Maguire

E. R.

(Eds.). (2007). Criminal justice theory: Explaining the nature and behavior of criminal justice. New York, NY: Routledge.

15.

Engel

R. S.

Calnon

J. M.

Bernard

T. J.

(2002). Theory and racial profiling: Shortcomings and future directions in research. Justice Quarterly, 19, 201–225.

16.

Flannery

Kretschmar

(2012). Fugitive safe surrender: Program description, initial findings, and policy implications. Criminology & Public Policy, 11, 437–459.

17.

Fletcher

(1990). What cops know. New York, NY: Simon & Schuster.

18.

Goffman

(2009). On the run: Wanted men in a Philadelphia ghetto. American Sociological Review, 74, 339–357.

19.

Goldkamp

J. S.

Vilcica

E. R.

(2008). Targeted enforcement and adverse system side effects: The generation of fugitives in Philadelphia. Criminology, 46, 371–409.

20.

Gottfredson

Hirschi

(1990). A general theory of crime. Stanford, CA: Stanford University Press.

21.

Gray

K. M.

Fields

Maxwell

S. R.

(2001). Examining probation violations: Who, what, when. Crime and Delinquency, 47, 537–557.

22.

Guy

Platt

Zwerling

Bullock

(1985). Mental health status of prisoners in an urban jail. Criminal Justice and Behavior, 12, 29–53.

23.

Guynes

Wolff

(2004). Un-served arrest warrants: An exploratory study. Washington, DC: National Institute of Justice.

24.

Hager

Daniel

Graycarek

Knowles

(2005). Improved coordination and information could reduce the backlog of un-served warrants. Frankfort, KY: Legislative Research Commission.

25.

Harris

D. A.

(1997). “Driving while black” and all other traffic offenses: The supreme court and pretextual traffic stops. Journal of Criminal Law and Criminology, 87, 544–582.

26.

Hartley

Madden

Spohn

(2007). Concerning conceptualization and operationalization: Sentencing data and the focal concerns perspective—A research note. The Southwest Journal of Criminal Justice, 4, 58–78.

27.

Higgins

Vito

Grossi

(2012). The impact of race on the police decision to search during a traffic stop: A focal concerns theory perspective. Journal of Contemporary Criminal Justice, 28, 166–183.

28.

James

D. J.

Glaze

L. E.

(2006). Mental health problems of prison and jail inmates (No. NCJ 213600). Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice, Office of Justice Programs.

29.

Johnson

(2003). Racial and ethnic disparities in sentencing departures across modes of conviction. Criminology, 41, 449–489.

30.

Johnson

B. D.

Ulmer

J. T.

Kramer

J. H.

(2008). The social context of guidelines circumvention: The case of federal district courts. Criminology, 46, 737–783.

31.

Karberg

James

(2005). Substance dependency, abuse, and treatment of jail inmates, 2002 (No. NCJ 209588). Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice, Office of Justice Programs.

32.

Klinger

D. A.

Bridges

(1997). Measurement error in calls-for-service as an indicator of crime. Criminology, 35, 705–726.

33.

Kraska

P. B.

Brent

J. J.

(2011). Theorizing criminal justice: Eight essential orientations. Long Grove, IL: Waveland.

34.

Maxfield

M. G.

Lewis

D. A.

Szoc

(1980). Producing official crimes: Verified crime reports as measures of police output. Social Science Quarterly, 61, 221–236.

35.

Miles

Shevlin

(2001). Applying regression and correlation. Thousand Oaks, CA: Sage.

36.

Moskos

(2008). Cop in the Hood. Princeton, NJ: Princeton University Press.

37.

Mumola

Karberg

(2006). Drug use and dependence, state and federal prisoners, 2004. (No. NCJ 21350). Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice, Office of Justice Programs.

38.

Pollock

(2012). Things change: An intergenerational examination of the correlates of police contact. Crime & Delinquency. Retrieved April 4, 2012, from http://cad.sagepub.com/content/early/2012/03/31/0011128712437913

39.

Pollock

Oliver

Menard

(2012). Measuring the problem: A national examination of disproportionate police contact in the United States. Criminal Justice Review, 37, 153–173.

40.

Sevigny

Caulkins

(2004). Kingpins or mules: An analysis of drug offenders incarcerated in federal and state prisons. Criminology & Public Policy, 3, 401–434.

41.

Shapiro

S. S.

Wilk

M. B.

(1965). An analysis of variance test for normality. Biometrika, 52, 591–599.

42.

Skolnick

(1966, 1994). Justice without trial. New York, NY: John Wiley.

43.

Smith

M. R.

Alpert

G. P.

(2007). Explaining police bias: A theory of social conditioning and illusory correlation. Criminal Justice and Behavior, 34, 1262–1283.

44.

Smith

E. L.

Durose

D. R.

(2006). Characteristics of drivers stopped by the police, 2002. Washington, DC: Bureau of Justice Statistics.

45.

Steffensmeier

(1980). Assessing the impact of the women’s movement on sex-based differences in the handling of adult criminal defendants. Crime and Delinquency, 26, 344–357.

46.

Steffensmeier

Kramer

Streifel

(1993). Gender and imprisonment decisions. Criminology, 31, 411–446.

47.

Steffensmeier

Ulmer

Kramer

(1998). The interaction of race, gender, and age in criminal sentencing: The punishment cost of being young, black, and male. Criminology, 36, 763–797.

48.

Therneau

T. M.

Grambsch

P. M.

(2000). Modeling survival data: Extending the Cox model. New York, NY: Springer.

49.

Tillyer

Hartley

(2010). Driving racial profiling research forward: Learning lessons from sentencing research. Journal of Criminal Justice, 38, 657–665.

50.

Ulmer

J. T.

Johnson

(2004). Sentencing in context: A multilevel analysis. Criminology, 42, 137–177.

51.

U.S. General Accounting Office. (2001). Welfare reform: Implementation of fugitive felon provisions should be strengthened. Washington, DC: U.S. General Accounting Office.

52.

U.S. Office of the Inspector General Office of Audit. (2000). Identification of fugitives receiving supplemental security income payments. Washington, DC: Office of the Inspector General Office of Audit.

53.

U.S. Senate. (2001). Fugitives: The chronic threat to safety, law, and order. Washington, DC: U.S. Government Printing Office.

54.

U.S. Sentencing Commission. (2013). Guidelines manual. Retrieved from www.ussc.gov

55.

von Hirsch

(1976). Doing justice. Boston, MA: Northeastern University Press.

56.

Walters

S. J.

(2009). What is a Cox model? (2nd ed.). London, England: Hayward Medical Communications.

57.

Welte

Barnes

Hoffman

Wieczorek

Zhang

(2005). Substance involvement and the trajectory of criminal offending in young males. The American Journal of Drug and Alcohol Abuse, 31, 267–284.

58.

White

J. M.

(2000). Outstanding warrants in Milwaukee County: Fugitives from the justice system. Thiensville: Wisconsin Policy Research Institute.

59.

Wright

Decker

(1997). Armed robbers in action: Stickups and street culture. Boston, MA: Northeastern University Press.