Escapes From Correctional Custody

Abstract

Despite a shared interest in escapes from correctional custody by policy makers, facility administrators, media, and the public, there is a dearth of empirical research on this event. Our study synthesizes prior research, distinguishing between inmate-, incident-, and facility-level variables. We introduce the Correctional Incident Database, which employs a novel approach for collecting escape data. Finally, we describe a sample of 611 inmates involved in 503 escape incidents from 398 facilities. Our findings indicate that escapes from jail are a frequent, yet overlooked, phenomenon. The results also challenge the (mis)perception that escapes are often sensational and violent events.

Keywords

escape prison jail misconduct policy

Introduction

It has been said that “if there were such a thing as first principles in the field of corrections, the idea that prisons ought to prevent inmates from escaping would certainly qualify for the list” (Culp, 2005, p. 270). Several sectors of society, including prison administrators, the public, policy makers, and the media, share this notion regarding the importance of escapes. Correctional administrators, for example, consider escape risk to be an important predictor during inmate classification (Austin, 2003). Escapes also cause fear and pique curiosity in public (Culp, 2005; Fisher, Allan, & Allan, 2004), especially affecting those who live in close proximity to a prison (Carlson, 1990). Through legislation and judicial decisions, policy makers have created sentencing enhancements and long prison terms for individuals convicted of escape. Furthermore, escapes have been a popular topic in the media (Peterson, 2014), including fictional television shows, movies, and print and televised news.

Despite the popularity of escapes elsewhere in society, they have received little research attention from academia in the past 15 years. This lack of empirical study has led to common perceptions that escapes are sensational, often violent, correctional events (Peterson, 2014). While there have been two attempts to understand the context and outcomes of escapes (Culp, 2005; U.S. Sentencing Commission, 2008), these are both limited in breadth and depth. Thus, the current research has three goals. First, we seek to provide a thorough review of the research, culminating into a new conceptualization of escapes from custody. Next, we discuss the methodology of the Correctional Incident Database (CID), an innovative new database of escape incidents. Finally, we use these data to describe 611 inmates who escaped from custody in 2009, the circumstances of these incidents, and the characteristics of the facilities from which these escapes occurred. We end with a discussion of the implications for theory, policy, and future research.

Previous Escape Research

It is important to recognize the dearth of available escape research. Existing research is methodologically limited and outdated (Sturrock, Porporino, & Johnston, 1991). The most current and seminal research comes from Culp (2005) and the U.S. Sentencing Commission (2008). Culp analyzed several sources of national data to demonstrate the overall trend of escapes, as well as a small sample of media reports to understand more qualitative components of escape. The U.S. Sentencing Commission’s report examined the amount of violence associated with escapes. Both studies have been cited in U.S. Circuit Court decisions (e.g. U.S. v. Chambers, 2007; U.S. v. Templeton, 2008). Still, even these seminal studies are limited in scope. Culp’s primary analysis excluded walkaways, AWOLs (Absent Without Leave), escapes that occurred during transport, and escapes from jails, whereas the Sentencing Commission’s sample consisted of federal court cases only.

Another important limitation in the existing literature is the inconsistent definition of “escape.” Definitions can be found in, and vary across, legal codes, professional correctional organizations, and research. For example, while some do not distinguish between completed and attempted escapes (AL Penal Code § 13A-10-31; U.S. Sentencing Commission, 2008), the Association of State Correctional Administrators (ASCA) asserts that an escape from a facility has not occurred until “the inmate breaches the last line (barrier) of security” (ASCA, 2012, p. 12). Furthermore, while most previous research does not differentiate between different types of escapes, others have defined escapes from higher security facilities, walkaways from lower security facilities, failures to return from an authorized rerelease, and escapes during transport (ASCA, 2012; Culp, 2005; U.S. Sentencing Commission, 2008). The lack of consistency among definitions of escape erodes the construct validity of escape research (Culp, 2005), making it questionable to compare existing escapes statistics, research, policies, or sanctions. Despite these limitations, previous research has generally attempted to answer three important questions: who escapes?; from where?; and under what circumstances?

Who Escapes?

Much of the available scholarly work on escapes has attempted to explain why some inmates are more likely to escape from custody than other inmates, focusing mostly on the differences in demographics and criminal histories between escapees and nonescapees. Some of these findings are inconsistent. For example, some research has shown that men escape at higher rates than women (Chard-Wierschem, 1995; Lyons, 2011), but this is not substantiated in more recent, rigorous work (Culp, 2005). In addition, much of the previous literature indicates that White inmates tend to escape at higher rates than Black inmates (Cowles, 1981; Murphy, 1984; Sandhu, 1996), again, though more current research suggests that race is no longer an important predictor of escape (Culp, 2005), including failures to return (Chard-Wierschem, 1995) and walkaways (Johnston & Motiuk, 1992).

Researchers have also investigated the role of sentence length under the assumption that long sentences provide incentive for inmates to escape (Sturrock et al., 1991). Some studies have shown that escapees were serving longer sentences than nonescapees (Scott, Mount, & Duffy, 1977; Virginia Department of Corrections, 1978), whereas others reported that escapees often serve short sentences of 5 years or less (Morgan, 1967) or that sentence length is not a useful predictor of whether or not an inmate will escape (Holt, 1974).

It seems logical that inmates with more time left on their sentence are more likely to escape than inmates who have served most of their sentence. The research, though, is inconclusive. Earlier research (Hilbrand, 1969; McNeil, 1978) indicates inmates are more prone to escaping after serving only a small portion of their sentence, motivated to “escape as soon as possible in order to avoid a lengthy period of confinement” (Sturrock et al., 1991, p. 7). Culp and Bracco’s (2005) research, however, has demonstrated that some inmates escape even after they have served a sizable portion of their sentences. Thus, escaping from custody appears to be driven by both motivation and opportunity.

Other findings are consistent across studies. For example, younger inmates are more likely to escape from custody than older inmates (Anson & Hartnett, 1983; Lyons, 2011; Stone, 1975). The average age of escapees has been estimated at 25 (Morgan, 1967), 30 (Sturrock et al., 1991), and 34 (Culp, 2005) years. Younger inmates are more likely to engage in multiple types of escapes, including failures to return/AWOLs (Chard-Wierschem, 1995) and walkaways from minimum security facilities (Johnston & Motiuk, 1992).

The literature has also consistently shown that inmates convicted of property offenses are more likely to escape than inmates convicted of other crimes, such as violent offenses (Basu, 1983; Culp, 2005; Murphy, 1984; Thornton & Speirs, 1985). It has been estimated that escapees were 20% to 26% more likely to be serving time for a property crime compared with nonescapees (Holt, 1974; Murphy, 1984). Despite this apparent association between committing a property offense and escaping, “there seems to not be a clear explanation in the literature as to why this relationship persists” (Sturrock et al., 1991, p. 4).

Escapees also tend to have a history of escaping (Murphy, 1984; Sandhu, 1996; Thornton & Speirs, 1985) and are approximately twice as likely to have a history of escape when compared with a control group of nonescapees (Holt, 1974). Similarly, escapees have generally been incarcerated more frequently (Basu, 1983; Holt, 1974), have longer criminal histories (Murphy, 1984; Shaffer, Bluoin, & Pettigrew, 1985), more parole violations (Basu, 1983; Murphy, 1984), and more institutional violations (Murphy, 1984; Stone, 1975) than nonescapees.

Other research has examined more “dynamic” differences between escapees and nonescapees (Sturrock et al., 1991). For example, escape behavior is consistently linked to family problems, such as having problems with significant others (Basu, 1983; Sandhu, 1996), not receiving mail or personal visits from family members (Hilbrand, 1969; McNeil, 1978), or being placed in facilities far away from home (Kentucky Bureau of Corrections, 1979). Other dynamic factors associated with escape include “institutional crisis situations” which often involve conflict with guards or other inmates, such as the threat of physical or sexual assault (McNeil, 1978; Murphy, 1984; Sandhu, 1996). Administrative sanctions and decisions can also be catalysts to escape behavior, such as denial of parole (Kentucky Bureau of Corrections, 1979; Virginia Department of Corrections, 1980), delays in parole hearings (Holt, 1974), and transfers to higher security facilities (Murphy, 1984; Sandhu, 1996).

From Where?

In addition to distinguishing between escapees and nonescapees, researchers and correctional administrators have attempted to understand the facility-level characteristics that make prisons and jails more susceptible to escapes. The construction and technology of facilities, for instance, may influence the number of escapes that occur within a particular institution. Effective perimeter security can act as both a “psychological deterrence” and a “physical barrier” to escapes (McManus & Conner, 1994, p. 142). Another facility-level indicator that is often cited in previous research is overcrowding. While overcrowding has been found to influence behaviors such as infractions (Wooldredge, Griffin, & Pratt, 2001) and assaults (Gaes & McGuire, 1985), the true direction of the relationship between crowding and misbehavior is not well understood (Franklin, Franklin, & Pratt, 2006). Likewise, the relationship between overcrowding and prison escapes is unclear. Jan (1980) reported a moderate, negative relationship between the escape rate and crowding in a male adult prison (r = −.42). Conversely, Anson and Hartnett (1983) concluded that “overcrowding exhibits a surprisingly low relationship to institutional escapes and is an altogether unimportant predictor” (p. 40).

Prison privatization is another facility-level characteristic that can influence inmate behavior. Some scholars have argued that private correctional facilities have higher levels of total misconduct and violent misconduct than comparable public facilities (Camp, Gaes, Langan, & Saylor, 2003). Similarly, the findings of some research suggest that escapes are more likely to occur in private facilities than state-run facilities (Archambeault & Deis, 1998). Despite these concerns, Culp (2001) found that private prisons actually had statistically lower escape rates than public prisons. In addition, publicly operated facilities were more likely to experience multiple escapes compared with private prisons (Culp, 2001). Again, limited and outdated research makes it difficult to determine the real effect of the facility operator on escapes.

Anson and Hartnett (1983) conducted the most thorough examination into the effects of facility-level correlates of escape. In their analysis of adult male prisons in Georgia, they found that four variables—age of the inmate population, ratio of treatment staff to inmates, level of supervision, and annual resources per inmate—accounted for 70% of the variance in escape levels across facilities (Anson & Hartnett, 1983). Jan’s (1980) findings also suggest that facilities with youthful populations are more likely to experience escapes than adult facilities.

Under What Circumstances?

Some factors are situational and vary separately from the inmate or the facility. For example, an inmate might be held at a maximum security prison, but escape while in transport to an off-site medical facility. Thus, a situational or incident-level variable is whether inmates escape while inside or outside the facility in which they are confined. Culp and Bracco (2005) found that approximately 14% of the escapes reported by the media in 2001 occurred while the inmate was being transported between facilities. The U.S. Sentencing Commission (2008) estimated this number to be only 3%, but they did not distinguish between completed escapes and attempted escapes, making comparisons between these two studies difficult.

Another example of an incident-level variable is whether the escape occurred from a secure or nonsecure area. Both secure and nonsecure areas can be found inside or outside a facility. For example, secure areas typically have some type of barrier that the escapee must overcome. Inmates escaping from inside a facility must overcome perimeter security (e.g., fence, gate, walls, locked doors, etc.) for an escape to be from a secure area (see ASCA, 2012; U.S. Sentencing Commission, 2008). Escapes from within a facility can also be from nonsecure areas if escapees can leave without overcoming any barriers. These types of escapes are typically referred to as “walkaways” (U.S. Sentencing Commission, 2008). For incidents that occur outside a facility, escapees may be in a nonsecure area if they are on an authorized leave of absence, such as on furlough or work release. These types of escapes are often categorized as failures to return, AWOLs, or absconding. Escapes outside the facility could occur in a secure area if the inmate is being supervised by a staff member, such as during a custodial work detail, medical or court visit, or transportation.

Previous research has universally and historically demonstrated that escapes are most likely to occur in lower security institutions and other nonsecure settings. This has been attributed to the fact that less secure settings provide inmates with more opportunity to escape (Sturrock et al., 1991). Culp estimated that 88.5% of all prison escapes occurred in low security facilities. Findings from other studies indicate that inmates who are on work assignments have the best opportunity to escape (Duncan & Ellis, 1973; Holt, 1974; Kentucky Bureau of Corrections, 1979; Virginia Department of Corrections, 1978).

Still, not all escape methods involve taking advantage of opportunity. Some methods involve cutting through, climbing over, tunneling under, or otherwise defeating the perimeter of a facility. The frequency of these types of methods being employed has been estimated to occur in anywhere from 24.3% (Culp & Bracco, 2005) to 44.4% (Culp, 2005) of all escapes. Escapes could also include forging documentation, such as swapping identification or creating phony release papers, to trick correctional staff into releasing an inmate who is not supposed to be released (Culp, 2005). Even though the escape method is a useful incident-level factor that can provide insight into the escape process, research in this area is particularly limited. In the two studies identified that specifically examine escape method, the sample sizes consisted of only 70 (Culp & Bracco, 2005) and 72 (Culp, 2005) escapees.

Other incident-level findings suggest that escapes occur more frequently in warmer months, such as in spring or summer and on weekend days (Dahlem, 1974; Hilbrand, 1969; Kentucky Bureau of Corrections, 1979; McNeil, 1978; Murphy, 1984; Virginia Department of Corrections, 1978, 1980). Culp and Bracco (2005) also found that 70% of escapes occurred in either the morning (6:00 a.m.-12:00 p.m.) or evening (6:00 p.m.-12:00 a.m.) hours. A final incident-level variable that has not garnered much attention in the literature is the type of assistance inmates receive during the escape. Inmates can either receive inside assistance from a correctional staff member or outside assistance from a friend or family member helping the inmate escape from the facility and/or avoid recapture. Both types of assistance are rare (Culp, 2005; Culp & Bracco, 2005).

A New Conceptualization of Escapes

As evidenced in prior research, each escape from custody includes inmate-, incident-, and facility-level variables. These variables can be nested in a structure so that one facility may have multiple incidents, and each incident may involve several escapees (see Figure 1). Thus, an ideal examination of escapes from custody should examine the individual and interactive effects of these three levels of variables (Peterson, 2015). Also important, distinguishing between these levels of variables allows researchers to develop better theoretical models to explain escapes from custody. To date, previous research has not examined this nested structure.

Figure 1.

Hierarchical escape structure.

The CID

The CID uses data on recent escapes to build on this conceptualization of the escape process by compiling information on the escapee, the incident, and the facilities from which the escape occurred. The CID is the only available database on escape incidents from both prisons and jails across the United States. Furthermore, despite the fact that the CID only includes data from 2009, these data are much more recent, and thus more relevant, than data used in any other study of which we are aware. Finally, the CID uses a central definition of escape, which allows comparison of escapes to be made across jurisdictions and improves internal validity. Data for the CID come from an exhaustive, open-source search protocol adapted from the Extremist Crime Database (Freilich, Chermak, Belli, Gruenewald, & Parkin, 2014). Data were collected in a three-step protocol that included identifying escape incidents (sourcing), performing an exhaustive search for all information on that escape (searching), and coding the data for analysis (coding).

Sourcing

To identify escape incidents to be included in the study, project researchers used LexisNexis, correctional incident archives (such as corspecops.com), and state Department of Corrections websites. Potential cases were screened using specific inclusion criteria, which provided uniform results across jurisdictions where the definition of escape may vary. Inclusion criteria included that the escape happened in the United States, that the escapee be under some form of correctional custody, that there was an actual loss of control, and that it occurred in 2009.

Searching

Once an incident was identified and found to fit the inclusion criteria, data were collected on the facility, incident, and inmates associated with that escape. Using a hierarchy of trust, all available information was gathered from government publications, state Department of Corrections websites, newspapers, and other online resources. Intersearcher reliability checks were made to make sure that the searches conducted were exhaustive in nature.

Coding

Once information was gathered on each escape, the research team conducted a quantitative content analysis to code for facility, incident, and escapee variables. Facility-level information was also gathered from the American Correctional Association’s (ACA; 2010) Directory of Adult and Juvenile Correctional Departments, Institutions, Agencies, and Probation and Parole Authorities (71st Edition) and National Jail and Adult Detention Directory (12th Edition; ACA, 2012), as well as the Annual Survey of Jails: Jail-Level Data, 2010 (Bureau of Justice Statistics, 2011). The research team also supplemented escapee-level information with targeted searches on state Department of Corrections websites and the Victim Information Notification Everyday website, which sometimes yielded additional demographic and criminal history information. In cases of conflicting information, more credence was given to government sources than news media and more credence to larger news organizations than smaller ones. Also, the number of sources reporting the information was taken into account. As a result, the most reliable information was used in the database. There were intercoder reliability checks at regular intervals, and every case was reviewed by one of two senior coders (for more information on the CID, see Peterson, 2015).

Findings and Discussion

The CID includes data on 611 inmates who were involved in 503 escape incidents from 398 facilities in 2009. Thus, each facility in this sample experienced an average of 1.26 escapes, while each escape involved an average of 1.21 inmates. It is important to first note that a limitation of the CID is the amount of missing data. Many of the variables discussed below are missing data for a portion of the observations. In the case of the CID, missing data are often an artifact of the data collection method: Some information is not provided in the news articles and reports used for coding. We revisit this limitation in the final section of this study.

Table 1 presents descriptive statistics on the facilities from which inmates escaped. These facilities were generally undercapacity, with an average capacity level of approximately 96%. Most facilities were not accredited by the ACA, and the majority (68%) of facilities only held adults, although some held juveniles (3.9%), or both adults and juveniles (28.1%). The average rated capacity of these facilities was about 216 inmates, but this ranged from 3 to almost 2,000. The mean age of the facilities was 31 years, although some were built very recently and one was built in 1814. Furthermore, there was also much variation in terms of staff to inmate ratio. On average, there were almost five inmates per staff member at these facilities, but, again, there was substantial variation with facilities having more than 16 inmates per staff member, whereas other facilities had more staff than inmates.

Table 1.

Facility-Level Variables (n = 398).

Variable Name	Frequency		%
Facility level
Federal	13		3.3
State	141		35.5
Regional	8		2.0
County	216		54.4
Municipal	19		4.8
Security classification
Community corrections/work release	42		10.9
Minimum/low	68		17.7
Medium	26		6.8
Maximum	30		7.8
Jail/unclassified	218		56.8
Age classification
Adult	227		68.0
Juvenile	13		3.9
Mixed	94		28.1
Private prison
Yes	33		8.4
No	361		91.6
ACA accreditation
Yes	66		17.1
No	320		82.9
	M	SD	Minimum	Maximum
Age of the facility (years)	31.08	28.68	0	195
Size (rated capacity)	215.67	319.52	3	1,870
% capacity	95.99	29.51	11.54	211.61
Inmate to staff ratio	4.53	2.52	.27	16.42

Note. Missing cases are not shown; thus, the sum of the frequencies for all categories in each variable will not necessarily equal 398. ACA = American Correctional Association.

Jails accounted for a large proportion of the escapes recorded in the CID: More than 60% of the facilities which had an escape in 2009 were operated by local jurisdictions compared with fewer than 40% operated at the state or federal level. While this is not surprising in light of the number of inmates who cycle through jail each year, escapes from local facilities have generally been overlooked in previous literature.

As the majority of the escapes occurred in jails, a large proportion (56.8%) of facilities were not given a security classification.¹ Only 110 of the 398 facilities were classified as community corrections/work release or minimum security. Moreover, 26 facilities were classified as medium security, and 30 were classified as maximum security. Initially, these findings seem to contradict a well-support fact in previous escape research: The majority of escapes occur in low security facilities (see Culp, 2005). However, there are two differences between the current study and previous research that could explain this discrepancy. First, the current study includes escapes from jails, which have almost universally been excluded in previous research. If we exclude jail/unclassified facilities from the analysis, facilities classified as community corrections/work release or minimum/low comprise approximately 66% of the total number of facilities of where escapes occurred. The second explanation for this discrepancy stems from the CID’s open-source data collection. Escapes from higher security facilities are more likely to garner media attention (Peterson, 2014) and were thus more likely to be included in the CID.

Table 1 also indicates that only 8.4% of the facilities that experienced an escape in 2009 were privately operated. Recent numbers suggest that 23% of all state and federal prisons are privately operated (Stephan, 2008). However, these numbers do not include locally operated jails. In addition, this difference may not be important if other factors—such as the security classification of the facility—are not controlled for. Therefore, the role of prison privatization in preventing or permitting escapes remains unclear.

Descriptive statistics on the 503 escape incidents identified in the CID can be found in Table 2. Unlike previous research, which has generally only categorized incidents as either “escapes” or “walkaways,” the CID created a more detailed typology of escapes. A “force” escape occurred when at least one escapee used or threatened to use force against a staff member to facilitate the breakout. An escape was categorized as a “flight” when an inmate overcame some barrier to flee from custody, such as a fence, wall, window, or the direction observation of a staff member. “Collusions” occurred when a staff member was actively involved in helping the inmate(s) escape from custody. Inmates could have also used “fraud/deception” to facilitate their escape by assuming the identities of other inmates who were being released, dressing as civilians who were visiting the facility, or forging release documents. Some escapes simply involved inmates who “walked away” from facilities with no barriers to prevent them from leaving. These escapes typically occurred in facilities classified as community corrections/work release or minimum/low security. Finally, “failures to return/AWOLs” occurred when inmates were authorized to leave the facility—such as during work release or on a weekend furlough—and fail to return by the predetermined time.²

Table 2.

Incident-Level Variables (n = 503).

	Frequency	%
Escape type
Force	43	9.1
Flight	182	38.5
Collusion	6	1.3
Fraud/deception	27	5.7
Walkaway	156	33.0
Failures to return/AWOL	59	12.5
Assistance
Inside	3	0.6
Outside	73	15.6
Inside and outside	5	1.1
None	386	82.7
Plan
Yes	103	20.5
No (unknown)	398	79.4
Location
Inside	230	51.6
Outside	216	48.4
Time of incident
12:00 a.m.-6:00 a.m.	62	14.9
6:01 a.m.-12:00 a.m.	108	26.0
12:01 p.m.-6:00 p.m.	129	31.0
6:01 p.m.-11:59 p.m.	117	28.1
Day of week
Monday-Friday	360	73.2
Saturday-Sunday	132	26.8
Season
Winter (December, January, February)	112	22.4
Spring (March, April, May)	109	21.8
Summer (June, July, August)	142	28.3
Fall (September, October, November)	138	27.5

Note. Missing cases are not shown; thus, the sum of the frequencies for all categories in each variable will not necessarily equal 503.

As only 8.5% of escapes involved any force, it appears as though escapes are generally not associated with violence. Again we see that, compared with previous research, a surprisingly low percentage of escapes (45%) were classified as walkaways or AWOLs. This finding is also likely due to the large number of escapes from jails included in the current study. In addition, the CID is unique from previous research in its escape typology, especially the inclusion of “flight” escapes. In the past, researchers seem to have defined all escapes from low security as “walkaways,” despite the circumstances surrounding the incident. In the CID, if an inmate escaped from a low security facility, but had to climb over a wall or break through a window in the process, that incident would have been coded as a flight, not a walkaway. Conversely, if an inmate was being housed in a medium or maximum security facility, but escaped while on a weekend furlough, that escape would have been classified as an AWOL, whereas other research may not have been able to make that distinction. Therefore, even though definitional and data collection differences make it difficult to compare the findings from the current study with previous work, they do indicate the utility of obtaining detailed information about each escape incident. Specifically, this process allows one to better distinguish between different types of escapes.

Table 2 also indicates that a little more than 15% of incidents involved assistance from an individual outside the prison, such as a family member or friend of one of the escapees. Usually, this assistance came in the form of providing the escapee with a hiding place while he or she was out of custody. However, in some of the cases, outside assistance involved someone sneaking tools into the facility to facilitate the escape. There are very few incidents in which a staff member of the facility helped the inmate escape, but the level of help provided by staff members varied from driving the inmate away from the facility to assisting the escapee assault another staff member and break out of the facility. Still, a great majority of incidents (almost 83%) did not involve any assistance.

Interestingly, the number of incidents that occurred inside the facility (n = 230) was almost equal to the number of incidents that occurred outside the facility (n = 216). Escapes occurring outside the facility took place when inmates were being transferred, visiting an off-site medical or dental facility, or during an authorized release (e.g., furlough or work release). The fewest number of incidents (n = 62) occurred between 12:00 a.m. and 6:00 a.m., but escapes appear to occur during all hours of the day. Likewise, escapes appear to occur (proportionately) equally during all days of the week with a 73.2% likelihood of escapes occurring during between Monday and Friday and a 26.8% likelihood of occurring on the weekend. Finally, escapes were most likely to occur in the summer (28.3%) and fall (27.5%), which is partially consistent with previous research showing that criminal behavior, especially aggressive crime, increases in warmer months (Butke & Sheridan, 2010).

Another unique incident-level variable measured in the CID is whether the escape was planned. Escapes were coded as being planned if the escapee(s) involved in the incident told another inmate of their escape before it happened or wrote down their plan somewhere. Escapes were also coded as planned if the methods employed clearly involved forethought. For example, one of these incidents involved a long period of time spent sawing through the bars on the inmates’ cell window after one of the escapee’s wife smuggled a saw blade into the prison. Similarly, if someone was waiting in a car near the prison to help an inmate immediately after he or she escaped, that incident was seen as being premeditated. Approximately 20% of the escapes coded for this variable were seen as having been planned. Although this percentage may seem high, previous estimates of how many escapes are planned range from 40% (Sandhu, 1996) for any planning to 3% (Culp & Bracco, 2005) or 8% (Culp, 2005) for more sophisticated planning. It is also important to note that these “plans” are often very simplistic (see also Centre for Research, Evaluation, and Social Assessment, 1996; Culp, 2005; Culp & Bracco, 2005).

The final table presents the descriptive statistics on the 611 inmates who escaped in 2009. Consistent with previous research, escapees in the CID tend to be young, with an average age of approximately 30 years. Escapees are also overwhelmingly male, with fewer than 4% of escapees being female. In contrast, females made up approximately 7% of the total prison population and 12% of the total jail population in 2009 (Minton, 2010; West, Sabol, & Greenman, 2010). Furthermore, while more recent research has found that race does not predict escape behavior (Culp, 2005), findings from the current study seem to be more consistent with older research, indicating that White inmates are more likely to escape than Black inmates (e.g., Cowles, 1981; Holt, 1974; Morgan, 1967; Murphy, 1984; Sandhu, 1996; Stone, 1975; Virginia Department of Corrections, 1978, 1980, 1982). Table 3 indicates that nearly 60% of escapees were White, whereas only 34% of prisoners and 43% of jail inmates in 2009 were White (Minton, 2010; West et al., 2010).

Table 3.

Inmate-Level Variables (n = 611).

	Frequency		%
Sex
Male	583		96.4
Female	22		3.6
Race/ethnicity
White	292		56.6
Black	139		26.9
Hispanic	69		13.4
Other	16		3.1
Committing offense
Murder/attempted murder	45		9.7
Rape/sexual assault	17		3.7
Armed robbery/robbery	30		6.5
Burglary	84		18.1
Theft/larceny/forgery	84		18.1
Drug offense	94		20.3
Public-order offense	20		4.3
Probation/parole violation	37		8.0
Other	53		11.4
History of escape
Yes	51		11.0
No	414		89.0
Catalyst
Yes	37		6.8
No	507		93.2
Recaptured
Yes	559		91.5
No	52		8.5
	M	SD	Minimum	Maximum
Age	29.9792	9.84994	14	66
Sentence length (months)	156.80	221.77	0	1,451.50
Time left to serve (months)	103.64	214.03	0	1,573.00

Note. Missing cases are not shown; thus, the sum of the frequencies for all categories in each variable will not necessarily equal 611.

Given the proportion of medium and maximum security facilities identified in the CID that experienced escapes, it is not surprising that approximately 20% of the escapees in this database were convicted of or charged with violent crimes such as murder/attempted murder, rape/sexual assault, and armed robbery/robbery. Furthermore, almost 60% of escapees were serving time or awaiting trial for a drug or a property offense, and about 12% were in custody for a public-order offense or a parole/probation violation. For about 11% of the escapees in this database, this was at least their second escape from prison or jail.

On average, inmates were serving long sentences of approximately 13 years and had more than 8 years left on their sentence.³ However, there was significant variation in both of these variables. Given the stiff penalties and new criminal charges inmates face for escaping, it is notable that some inmates escaped while serving short sentences or while only having a short time left before they were to be released. One particularly perplexing—albeit rare—example from the CID involved an individual who was several days into a 30-day jail sentence, escaped from custody, was recaptured, and subsequently received a 2-year prison sentence for the escape. These findings highlight the opportunistic nature of many escapes.

We also see in Table 3 that a number of inmates escaped due to some catalyst event. Previous research indicates that these catalyst events often involve an administrative sanction or otherwise increased justice involvement, such as a parole denial, a transfer to higher security facility, or a charge for a new offense (Kentucky Bureau of Corrections, 1979; Murphy, 1984; Sandhu, 1996; Virginia Department of Corrections, 1980; Wharry, 1972). Consistent with these findings, catalyst events recorded in the CID include actions such as additional arrests while on furlough, threatened deportation, and pending new indictments/internal punishments. Likewise, many of the jail inmates escaped upon being convicted and/or finding out the length of their sentences. Other catalysts include the recent birth of one inmate’s daughter and being threatened by other inmates.

These findings also suggest that more than 92% of all inmates are captured after they escape. In one of the only other studies on recapture, Culp (2005) estimated the recapture rate to be around 74.5%. However, these findings were tempered by the fact that inmates escaping from higher security facility were more likely to be recaptured than inmates escaping from lower security facilities. Still, the difference between these two estimates could be a result of the data being used, the inclusion of jails in the current study, the differences in follow-up time, or the nearly decade difference between the two studies’ data collection periods.

Conclusions and Implications

The purpose of the current study was to establish a new conceptualization of escapes from custody and provide descriptive information relevant to this conceptualization. Nevertheless, there are several limitations to this research. First, data for the CID were collected using an open-source search protocol. While this methodology provides very detailed information, it also creates the potential for bias. For example, using a similar (although less thorough) methodology, Culp (2005) estimated that print media only reported on 6% of the escapes that occurred in 1997 and 9.2% of the escapes that occurred in 1998. Moreover, research has shown that media are more likely to report on violent crimes than nonviolent ones (e.g., Beckett, 1997; Marsh, 1991). Thus, it is likely that the escapes included in the CID are more violent or otherwise sensational than the typical escape (Peterson, 2014). Still, we believe it is important to point out that a rigorous effort was made to collect information from local news media and the official press releases of correctional agencies to mitigate some of this bias. Moreover, while it would have been desirable to have a more representative sample of escapes, escape data from prisons and jails are not currently being systematically collected.

As we previously discussed, another limitation with our method of data collection was that it led to much missing data. We did not have complete information on many of the variables examined in the previous section. This could have also led to biased interpretations because the data are likely not missing at random. It would be useful for future research to examine the sources of missingness in some of these variables to determine whether and to what degree missing data are influencing our interpretations of the data.

A final limitation of our study is that our data only include escapes from 2009. While these data are a few years old, we used the CID because it is the only existing data set with detailed information on escapes. Furthermore, the data collection for the CID was very labor-intensive, making it difficult to systematically collect these data over multiple years. Still, it is important to note that the results of our research are not generalizable to other periods of time.

This research has important implications for theory and policy. One finding in the current study is that many escapes appear to be opportunistic, supporting situational crime prevention theory. Even when escapes are planned, these plans are not very complex, and, accordingly, most escapes end in recapture. In applying the situational crime prevention theory to escapes, Wortley (2002) examined how much of their sentence inmates served before they escaped, the season and time of day in which inmates escaped, the institutional security level of the facility from which inmates escaped, and the location of the escape (i.e., inside or outside of the facility). These variables are already included in the CID and can improve the field’s understanding of situational crime prevention as it applies to escapes (see also Peterson, 2015).

In addition to these theoretical implications, findings from the current study inform policy decisions and discussions. Wortley’s (2002) research included strategies for controlling escapes, such as modifying the environment (e.g., increasing perimeter security), providing counseling to inmates, allowing more home visits and furloughs, offering more programming in the prison, and protecting inmates when their safety is threatened. Findings from the current study and previous research suggest that an important precursor to escaping is increased justice involvement. It might be prudent, then, for policy makers to focus counseling and supervision resources to inmates after they are sentenced, receive word of a transfer, and so on, to help mitigate the impact of this news on escape behavior. In addition, although granting more home visits and furloughs could provide more opportunity to escape, research has demonstrated that increased family contact can reduce institutional misconduct (Cochran, 2012; Siennick, Mears, & Bales, 2013).

A common perception of escapes—held by the public, as well as prison administrators, policy makers, and media—is that escapes are inherently violent (Peterson, 2015). This perception has directly influenced policy and sentencing decisions. The Armed Career Criminal Act of 1984, for example, is a sentencing enhancement structure requiring a minimum 15-year sentence for individuals with three prior violent felony convictions who are subsequently convicted of a weapons offense. According to the U.S. Circuit and Supreme Court’s most recent decisions, any escape more serious than a walkaway (e.g., scaling a fence) counts as a violent crime for purposes of a sentencing enhancement (see U.S. v. Stout, 2013). Our research challenges these assumptions: Even though our data collection process likely biased our sample toward more violence than what exists in reality, we demonstrate that violent methods (i.e., “force” escapes) are used in fewer than 10% of escape incidents. By routinely collecting data on escapes, this research has the potential to shed light on this issue. As a result, policy makers would be better equipped to make decisions based on data and not on common (mis)perceptions.

Finally, a critical discovery from the current study is that a high proportion of escapes are occurring in jails. In addition to the mixed population of pretrial detainees and sentenced inmates, jails tend to be independently operated and experience rapid population turnover (Minton, 2013). Furthermore, nearly 40% of all jail facilities hold fewer than 50 inmates (Stephan & Walsh, 2011). Although it was beyond the current study to identify the factors that might make jails prone to escapes, it is possible that this environment presents unique challenges for jail administrators and could make these facilities susceptible to escape. Just as the jail environment is different from the prison environment, policies for preventing escape should be tailored for each type of correctional facility.

Footnotes

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by a grant from the Academy for Critical Incident Analysis and from the Professional Staff Congress–City University of New York Research Award Program.

Notes

Author Biographies

Bryce Elling Peterson is a research associate in the Urban Institute’s Policy Center. His research focuses on correctional policy, technological interventions in criminal justice settings, and program evaluation.

Adam Fera is a doctoral student at the Graduate Center of the City University of New York (CUNY)/John Jay College of Criminal Justice. He is a presidential research fellow and adjunct lecturer as well as coproject coordinator of the Correctional Incident Database (CID).

Jeff Mellow is professor of criminal justice at John Jay College. His research interests include program implementation and evaluation, critical incidents in corrections, improving the health care needs of justice-involved individuals, and community corrections.

References

American Correctional Association. (2010). Directory: Adult and juvenile correctional departments, institutions, agencies, and probation and parole authorities (71st ed.). Alexandria, VA: Author.

American Correctional Association. (2012). 2011-2013 national jail and adult detention directory (12th ed.). Alexandria, VA: Author.

Anson

R. H.

Hartnett

C. M.

(1983). Correlates of escape: A preliminary assessment of Georgia prisons. Criminal Justice Review, 8(1), 38-42.

Archambeault

W. G.

Deis

D. R.

(1998). Cost effectiveness comparisons of private versus public prisons in Louisiana: A comprehensive analysis of Allen, Avoyelles, and Winn Correctional Centers. Baton Rouge: Louisiana State University.

Association of State Correctional Administrators. (2012). Performance-based measures system counting rules. Hagerstown, MD: Author.

Austin

(2003). Findings in prison classification and risk assessment. Washington, DC: U.S. Department of Justice, Federal Bureau of Prisons, National Institute of Corrections.

Basu

(1983). Models of escape behaviour of minimum security inmates in Ontario: The use of routine administrative data for risk assessments (Doctoral dissertation). Available from ProQuest Dissertations and Theses Full Text. (UMI No. 8309525)

Beckett

(1997). Making crime pay. New York, NY: Oxford University Press.

Bureau of Justice Statistics. (2011). Annual Survey of Jails: Jail-Level Data [Computer file]. Conducted by U.S. Department of Commerce, Bureau of the Census. ICPSR 33722-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research.

10.

Butke

Sheridan

S. C.

(2010). An analysis of the relationship between weather and aggressive crime in Cleveland, Ohio. Weather, Climate, and Society, 2, 127-139.

11.

Camp

S. D.

Gaes

G. G.

Langan

N. P.

Saylor

W. G.

(2003). The influence of prisons on inmate misconduct: A multilevel investigation. Justice Quarterly, 20, 501-533.

12.

Carlson

K. A.

(1990). Prison escapes and community consequences: Results of a case study. Federal Probation, 54(2), 36-43.

13.

Centre for Research, Evaluation, and Social Assessment. (1996). Escape pressures: Inside views of the reasons for prison escapes. Wellington: New Zealand Ministry of Justice.

14.

Chard-Wierschem

D. J.

(1995). Comparison of temporary release absconders and nonabsconders: 1993-1994. Albany: New York State Department of Correctional Services.

15.

Cochran

J. C.

(2012). The ties that bind or the ties that break: Examining the relationship between visitation and prisoner misconduct. Journal of Criminal Justice, 40, 433-440.

16.

Cowles

E. L.

(1981). A study to explore race as it relates to correctional institution escape behavior. Ann Arbor, MI: University Microfilms International.

17.

Culp

R. F.

(2001). A comparative analysis of escapes form public and private correctional facilities (Doctoral dissertation). Available from ProQuest Dissertations and Theses Full Text. (UMI No. 3008817)

18.

Culp

R. F.

(2005). Frequency and characteristics of prison escapes in the United States: An analysis of national data. The Prison Journal, 85, 270-291.

19.

Culp

R. F.

Bracco

(2005). Examining prison escapes and routine activities theory. Corrections Compendium, 30(3), 1-5, 25-27.

20.

Dahlem

G. G.

(1974). Escapes and escapees at Colorado State Penitentiary: Some speculations and statistics. Colorado Springs: Colorado Department of Corrections.

21.

Duncan

D. F.

Ellis

T. R.

(1973). Situational variables associated with prison escapes. American Journal of Correction, 35(3), 29-30.

22.

Fisher

Allan

M. M.

(2004). Exploratory study to examine the impact of television reports of prison escapes on fear of crime, operationalised as state anxiety. Australian Journal of Psychology, 56, 181-190.

23.

Franklin

T. W.

Franklin

C. A.

Pratt

T. C.

(2006). Examining the empirical relationship between prison crowding and inmate misconduct: A meta-analysis of conflicting research results. Journal of Criminal Justice, 32, 401-412.

24.

Freilich

J. D.

Chermak

S. M.

Belli

Gruenewald

Parkin

W. S.

(2014). Introducing the United States Extremist Crime Database (ECDB). Terrorism and Political Violence, 26(2), 372-384.

25.

Gaes

G. G.

McGuire

W. J.

(1985). Prison violence: The contribution of crowding versus other determinants of prison assault rates. Journal of Research in Crime & Delinquency, 22, 41-65.

26.

Hilbrand

R. J.

(1969). The anatomy of escape. Federal Probation, 33(1), 58-66.

27.

Holt

(1974). Escape from custody (Research Report No. 52). Sacramento: California Department of Corrections.

28.

Jan

J. L.

(1980). Overcrowding and inmate behavior: Some preliminary findings. Criminal Justice and Behavior, 7, 293-301.

29.

Johnston

J. C.

Motiuk

L. L.

(1992). Factors related to unlawful walkaways from minimum security institutions (Research Report No. R-23). Ottawa, Ontario: Correctional Service Canada.

30.

Kentucky Bureau of Corrections. (1979). The Blackburn escape study. Final report (NCJ No. 067405). Frankfort, KY: Commonwealth of Kentucky Department of Justice, Bureau of Corrections, Research and Evaluation Section. Retrieved from https://www.ncjrs.gov/pdffiles1/Digitization/67656NCJRS.pd

31.

Lyons

J. A.

(2011). Inmate escape incidents: 2006-2010. Albany: State of New York, Department of Correctional Services.

32.

Marsh

H. L.

(1991). A comparative analysis of crime coverage in newspapers in the United States and other countries from 1960–1989: A review of the literature. Journal of Criminal Justice, 19, 67-80.

33.

McManus

L. F.

Jr. Conner

J. C.

(1994, July). Deterring escapees through comprehensive perimeter security. Corrections Today, 56(4), 142-146.

34.

McNeil

(1978). Multivariate study of dynamic and static factors associated with prison escapes (Doctoral dissertation). Available from ProQuest Dissertations and Theses Full Text. (UMI No. 7811620)

35.

Minton

T. D.

(2010). Jail inmates at midyear 2009—Statistical tables. Washington, DC: Bureau of Justice Statistics.

36.

Minton

T. D.

(2013). Jail inmates at midyear 2012—Statistical tables. Washington, DC: Bureau of Justice Statistics.

37.

Morgan

(1967, March-April). Individual and situational factors related to prison escape. American Journal of Correction, pp. 30-31.

38.

Murphy

T. H.

(1984). Prediction of minimum security walkaways. Lansing: Michigan Department of Corrections.

39.

Peterson

B. E.

(2014). Newsworthiness of prison escapes: Content analysis of factors influencing print media coverage, 2006-2010. American International Journal of Social Science, 3(1), 174-187.

40.

Peterson

B. E.

(2015). Inmate-, incident-, and facility-level factors associated with escapes from custody and violent outcomes (Doctoral dissertation). Retrieved from http://academicworks.cuny.edu/gc_etds/606/

41.

Sandhu

H. S.

(1996). A profile of murderer escapees. Journal of the Oklahoma Criminal Justice Research Consortium, 3, 115-122.

42.

Scott

N. A.

Mount

M. K.

Duffy

P. S.

(1977). MMPI and demographic correlates and predictors of female prison escape. Criminal Justice and Behavior, 4(3), 285-300.

43.

Shaffer

C. E.

Bluoin

Pettigrew

C. G.

(1985). Assessment of prison escape risk. Journal of Police and Criminal Psychology, 1(1), 42-48.

44.

Siennick

S. E.

Mears

D. P.

Bales

W. D.

(2013). Here and gone: Anticipation and separation effects of prison visits on inmate infractions. Journal of Research in Crime & Delinquency, 50(3), 417-444.

45.

Stephan

J. J.

(2008). Census of state and federal correctional facilities, 2005. Washington, DC: Bureau of Justice Statistics.

46.

Stephan

J. J.

Walsh

(2011). Census of jail facilities, 2006. Washington, DC: Bureau of Justice Statistics.

47.

Stone

(1975). Factors related to escape prediction (Doctoral dissertation). Available from ProQuest Dissertations and Theses Full Text. (UMI No. 7811485)

48.

Sturrock

R. C.

Porporino

F. J.

Johnston

J. C.

(1991). Literature review on the factors related to escape from correctional institutions. Ottawa, Ontario: Research and Statistics Branch, Correctional Service Canada.

49.

Thornton

Speirs

(1985). Predicting absconding from young offenders institutions. In Farrington

D. P.

Tarling

(Eds.), Prediction in criminology (pp. 119-134). Albany: State University of New York Press.

50.

U.S. Sentencing Commission. (2008). Report on federal escape offenses in fiscal years 2006 and 2007. Washington, DC: Author.

51.

Virginia Department of Corrections. (1978). Report of felon escapes. Richmond, VA: Virginia Department of Corrections, Division of Program Development and Evaluation and Reporting Unit.

52.

Virginia Department of Corrections. (1980). Report of felon escapes. Richmond, VA: Virginia Department of Corrections, Division of Program Development and Evaluation and Reporting Unit.

53.

Virginia Department of Corrections. (1982). Report of felon escapes. Richmond, VA: Virginia Department of Corrections, Division of Program Development and Evaluation and Reporting Unit.

54.

West

H. C.

Sabol

W. J.

Greenman

S. J.

(2010). Prisoners in 2009. Washington, DC: Bureau of Justice Statistics.

55.

Wharry

J. D.

(1972). A study of the nature and frequency of crimes committed by escapees of maximum and medium security institutions and some characteristics of escapes. Ottawa, Ontario: Canadian Penitentiary Services.

56.

Wooldredge

Griffin

Pratt

(2001). Considering hierarchical models for research on inmate behavior: Predicting misconduct with multilevel data. Justice Quarterly, 18(1), 203-231.

57.

Wortley

R. K.

(2002). Situational prison control: Crime prevention in correctional institutions. Cambridge, UK: Cambridge University Press.