Gender,Co-Offending,and Recidivism among a Sample of Robbery and Burglary Offenders

Abstract

This study examines whether several aspects of co-offending are related to recidivism, and whether those relationships are gendered. A sample of 400 people (200 men and 200 women) released from prison after serving sentences for burglary or robbery is used to answer these questions. Results of Cox regression models showed risk of rearrest was lower for those who co-offended, those with more co-offenders, and those who co-offended with romantic partners or family members, while risk of rearrest was higher among those with leadership roles in the offense. However, gender-specific analyses revealed these relationships were only observed among women. The results point to the importance of gender in understanding recidivism and provide insights into effective correctional programming for women.

Keywords

co-offending gender recidivism

Introduction

Co-offending, or the perpetration of an offense by more than one person, is common among the majority of delinquent youth and a substantial portion of adult offenders (Reiss & Farrington, 1991; Warr, 1998, 2002) and is an important avenue for understanding crime. Violence occurs more often in offenses committed by groups than in those committed by individuals acting alone (McCord & Conway, 2005) and individuals who co-offend commit more crimes than solo offenders (Hindelang, 1976; Reiss & Farrington, 1991). Some scholars argue those involved in co-offending are more likely to commit subsequent offenses, especially if their co-offenders are violent (e.g., Andersen, 2019; Carrington, 2009; Conway & McCord, 2002; Lantz & Hutchison, 2015; Warr, 2002). However, other scholars have argued that recidivism is not related to co-offending or that any relationship that exists is not straightforward (e.g., Glueck & Glueck, 1950; Kornhauser, 1978; Ouellet et al., 2013). Given the mixed literature on the association between co-offending and recidivism and the importance of replication for developing knowledge (e.g., McNeeley & Warner, 2015), it is important to continue examining how co-offending might lead to further participation in crime.

Anti-social peers may be an especially important consideration for women. In particular, co-offending appears to be relevant for female crime (see van Mastrigt & Farrington, 2009), as women often get pulled into crime by male co-offenders (Becker & McCorkel, 2011; Mullins & Wright, 2003), although some studies suggest that men are equally or more likely to commit group offenses (McNeeley, 2019; Rennison, 2009). Women are more likely to commit serious, violent offenses when they work with men (Alarid et al., 1996; Becker & McCorkel, 2011; Koons-Witt & Schram, 2003). Recent research suggests that the nature of co-offending—such as one’s role in a particular offense and the relationship to co-offenders—varies by gender (McNeeley, 2019). Despite the literature on the gendered nature of co-offending, it is as yet unknown whether the relationship between co-offending and recidivism is gendered.

The current study examines (1) whether those who commit co-offenses are more or less likely than solo offenders to recidivate, (2) whether the nature of co-offending (such as the size of co-offending groups, the role in the offense, and the relationship with co-offenders) is associated with recidivism, and (3) whether these relationships vary by gender. To answer these questions, this study uses a sample of 400 people (200 males and 200 females) released from Minnesota state prisons after serving sentences for burglary or robbery. According to prior research, co-offending is especially important for understanding burglary and robbery, as they have higher rates of co-offending than other types of offenses (Carrington, 2002, 2009; Piquero et al., 2007; Reiss & Farrington, 1991; van Mastrigt, 2008; van Mastrigt & Farrington, 2009). Additionally, although burglary and robbery are predominantly committed by men (according to Pattavina et al., 2017, 14% of robbery arrestees and 19% of burglary arrestees in 2016 were women), these crime types have high rates of co-offending among both males and females (Carrington, 2009), making them ideal for analyzing the gendered nature of co-offending.

Co-Offending and Recidivism

Criminologists have studied how co-offending fits in to the criminal career, often suggesting that it increases the likelihood of recidivism. Co-offending is believed to increase social criminal capital (Lin, 2001; McCarthy & Hagan, 1995, 2001; Sutherland, 1947), as those who co-offend are parts of social groups centered around crime. After release from prison, people embedded in these social groups are more likely to be recruited to participate in specific criminal acts or may be able to seek out others for assistance in offenses of their own design. These opportunities can all increase the likelihood of recidivism. Additionally, after initially committing a crime, individuals’ cost-benefit analyses that structure their decisions to engage in later crimes can change based on past criminal experiences, making recidivism either more or less likely (see Cornish & Clarke, 1986). Co-offending can offer several rewards that influence this; those who co-offend may be less likely to be arrested and may earn larger amounts of money or goods (e.g., Bouchard & Nguyen, 2010; Lantz, 2020a; Morselli et al., 2006). Finally, participating in criminal acts with others may also provide less-tangible social rewards, such as the approval of influential peers. These can reinforce criminal behavior and make future crime more likely, in line with social learning theory (e.g., Akers, 1998). In line with these theories, some studies have found that subsequent offending behavior is higher among those who co-offend and that those who work with co-offenders have longer criminal careers (Andersen, 2019; Carrington, 2009; Conway & McCord, 2002; Lantz & Hutchison, 2015; McCord & Conway, 2002; McGloin & Piquero, 2009; Ouellet et al., 2013).

On the other hand, prior research has demonstrated a nuanced relationship between co-offending and recidivism. Ouellet et al. (2013) found that rearrest was less likely when the present offense was committed with a group, but that those with more co-offenders in their histories were more likely to be rearrested. Co-offending may in some circumstances be related to lower risk for recidivism, as co-offending is a situational process that may influence someone with low motivation to participate in a crime that they would not have committed alone (e.g., McGloin & Piquero, 2009; Ouellet et al., 2013; Warr, 2002). McGloin et al. (2008) suggest that many individuals tend not to engage in multiple offenses with the same co-offenders; therefore, co-offending may not be a strong signal that a person is embedded in criminal groups that make reoffending more likely. Some scholars maintain there is no relationship between co-offending and recidivism; rather, there is simply a selection effect (Glueck & Glueck, 1950; Kornhauser, 1978). These scholars argue group offenses often occur because a small number of offenders who are highly involved in crime are joined by co-offenders less likely to reoffend. Given these mixed findings, it is important to continue to examine the relationship between co-offending and recidivism. Hypothesis 1: Co-offending is related to recidivism, although the direction of the relationship is unclear.

Crucially, cooperation with others during a crime may not always have the same effect; rather, it is possible that the nature of co-offending is more important in understanding how it influences recidivism. For example, the number of co-offenders who cooperated in the offense may be related to recidivism, although the direction of the expected relationship is unclear. On the one hand, it may better represent the extent to which one has access to criminal capital and is therefore likely to recidivate (Lin, 2001; McCarthy & Hagan, 1995, 2001; Sutherland, 1947). On the other hand, it could indicate that the behavior was more greatly influenced by others and is even less likely to be repeated (e.g., McGloin & Piquero, 2009; Ouellet et al., 2013). Hypothesis 2: Recidivism is related to the number of co-offenders.

Next, the role one has in the commission of the offense may matter. Those who instigate co-offenses—play a major role in the planning and implementation of a crime event (see McGloin & Nguyen, 2012; Reiss, 1986)—may be likely to recidivate due to their individual high risk, while those who act as accomplices may have lower risk and therefore be less likely to commit new offenses. Further, accomplices may reap fewer financial and other rewards from the crime, making them less likely to recidivate than instigators. Hypothesis 3: Those with higher roles in a co-offense are more likely to recidivate. Finally, the relationship to the co-offenders may influence the extent to which someone is likely to engage in subsequent crime. Some relationships (e.g., between family members or romantic partners) are more persistent than others and as such may be a better indicator of criminal capital, making future co-offenses likely. Hypothesis 4: Those who co-offend with family members or romantic partners are more likely to recidivate.

Gender, Co-Offending, and Recidivism

Prior research suggests potential gender differences in the relationship between co-offending and recidivism. On one hand, co-offending may have criminogenic influences that are stronger among women than men. Women tend to commit crimes that are more serious and more violent in nature when they work with men than when they commit crimes alone (Alarid et al., 1996; Becker & McCorkel, 2011; Koons-Witt & Schram, 2003). This introduction to more serious crime may increase recidivism risk among women. Additionally, some studies suggest that deviant peers are especially criminogenic for women (Park et al., 2010), although other studies suggest the opposite (Crosnoe et al., 2002; Wall et al., 1993). Prior qualitative research suggests women who co-offend often do so with people they are close to, such as family members or romantic partners (Becker & McCorkel, 2011; Mullins & Wright, 2003). A recent quantitative study confirms that this relationship between co-offenders is more common among women than among men (McNeeley, 2019). Due to this close nature of the relationship between women and their co-offenders, co-offending may be a stronger measure of deviant peers among women than among men. Therefore, women who co-offend may be more likely to reoffend, especially if their relationship with the co-offender(s) continues after their release from prison.

On the other hand, women who co-offend may be less likely to continue participating in crime than men who co-offend. For example, prior research on female sex offenders found those who were incarcerated for co-offenses were less likely to recidivate than those who committed the offense alone (Muskens et al., 2011). Theoretically, we may expect co-offending to lead to lower recidivism among women for several reasons. First, as mentioned above, women who co-offend are more likely to commit crimes with romantic partners (Becker & McCorkel, 2011; McNeeley, 2019; Mullins & Wright, 2003). While those relationships can persist and make future criminal behavior more likely, they can also dissolve and make desistance seem possible or even ideal. Indeed, women released from prison often consciously choose to cut ties with antisocial friends or family members in order to increase their chances for success during reentry (Berg & Cobbina, 2017).

Second, within criminal networks, women are often not considered high-ranking or even full members, making them less able to plan and implement their own serious solo offenses or recruit others to join them in criminal acts (Schwartz et al., 2015; Schwartz & Steffensmeier, 2017; Steffensmeier et al., 2013). Instead, prior research suggests women who co-offend often act as accomplices in the offense, acting as lookouts, obtaining information, or attracting or distracting potential targets (McCarthy & Hagan, 1995; McNeeley, 2019; Miller, 1998; Mullins & Wright, 2003; Schwartz & Steffensmeier, 2017). Because these women are otherwise low risk, their choice to become involved in future crime likely depends more greatly on the rewards gained through these offenses (e.g., Cornish & Clarke, 1986). Since they act as accomplices and are not full members of criminal groups, they may receive fewer tangible rewards and their participation may not be strongly valued, so their co-offending may not provide the same level of reinforcement as that experienced by men. Therefore, women’s experience working with co-offenders may have less of a lasting effect than might be seen with men. Hypothesis 5: The relationships between recidivism and co-offending, the number of co-offenders, the role in the offense, and the relationship to the co-offender vary by gender, although the expected direction is unknown.

Research Methods

Data and Sample

This study examines a sample of burglary or robbery offenses committed by people released from Minnesota state prisons. First, 400 individuals (200 males and 200 females) released between January 2014 and April 2018 were selected. The sample includes all 200 women who were released on sentences for burglary and robbery during that time period¹ and a random selection of 200 men released during the same time period. Then, the burglary or robbery offenses for which they were incarcerated were examined. The 400 individuals in the sample were sentenced for 484 burglary or robbery offenses.

This study employs content analysis of pre-sentence investigation (PSI) reports and criminal complaints. PSIs are conducted by field service agents who interview the defendant as well as other individuals who can provide important information, such as police, prosecutors, mental health and substance abuse treatment providers, family members, associates, and employers. PSIs contain information on the circumstances of the offense, including the defendant’s version of the offense, as well as information about the defendant’s background. A criminal complaint is a written statement of the facts of the case and a description of the evidence against the defendant. Data collection was conducted using (1) the criminal complaint for the offense that was selected for the study and (2) the study subject’s PSI completed for the sentencing of that offense. In some cases when additional information was needed, data collection was also conducted using (3) the study subject’s PSIs completed for the sentencing of other offenses and/or (4) the PSI completed for the sentencing of the co-offender(s). All cases were coded by the author. Data on criminal offenses, including co-offending information, were collected through the criminal complaints and the descriptions of crime events provided in the PSIs. Individual data were collected from the Minnesota Department of Corrections’ (MnDOC) Correctional Operations Management System (COMS). Some PSI reports were incomplete, resulting in missing data, which was handled using multiple imputation.

Sample characteristics and descriptive statistics for all variables in the analyses are available in Table 1. Nearly half (49%) of the sample was White, while 30% was Black, 17% was Native American, 2% was Asian or Pacific Islander, and 2% was Hispanic. The sample ranged from 16 to 57 years of age, with an average age of 28 years. A majority of the offenses in the sample (70%) were burglaries. Compared to the general population of those incarcerated in Minnesota, this sample contains a higher percentage of Native Americans and is, on average, younger. For example, on January 1, 2018, 9.7% of the Minnesota prison population was Native American, and the average age was approximately 37 years.

Table 1.

Descriptive Statistics.

	Full sample			Males			Females			t
	Mean	SD	Min-max	Mean	SD	Min-max	Mean	SD	Min-max	t
Dependent variables
Rearrest	0.71	0.45	0–1	0.80	0.40	0–1	0.62	0.49	0–1	4.04*
Reconviction	0.58	0.50	0–1	0.65	0.48	0–1	0.51	0.50	0–1	2.96*
Reincarceration	0.23	0.42	0–1	0.29	0.45	0–1	0.17	0.37	0–1	2.90*
Revocation	0.45	0.50	0–1	0.51	0.50	0–1	0.39	0.49	0–1	2.43*
Independent variables
Any co-offense	0.61	0.49	0–1	0.60	0.49	0–1	0.62	0.49	0–1	−0.41
Average number of co-offenders	1.03	1.12	0–6	1.03	1.17	0–6	1.02	1.06	0–5	0.08
Highest role in an offense	2.46	0.72	1–3	2.60	0.62	1–3	2.33	0.78	1–3	3.83*
Any co-offense with romantic/family	0.21	0.41	0–1	0.14	0.35	0–1	0.28	0.45	0–1	−3.37*
Control variables
Robbery offense	0.30	0.46	0–1	0.35	0.48	0–1	0.26	0.44	0–1	2.07*
New commitment	0.84	0.37	0–1	0.82	0.39	0–1	0.87	0.34	0–1	−1.51
Metro commitment	0.48	0.50	0–1	0.55	0.50	0–1	0.41	0.49	0–1	2.93*
Length of stay (months)	35.50	21.38	0–129	39.34	23.04	0–129	31.65	18.85	0–120	3.65*
Discipline	6.39	13.78	0–121	6.68	13.47	0–103	6.10	14.11	0–121	0.42
Treatment	0.78	0.41	0–1	0.73	0.45	0–1	0.84	0.37	0–1	−2.56*
Visitation	16.47	45.87	0–500	17.30	45.59	0–386	15.65	46.25	0–500	0.36
Prior convictions	2.88	4.13	0–27	3.66	4.68	0–27	2.11	3.36	0–20	3.79*
LS/CMI	27.02	6.63	7–41	27.45	6.91	11–40	26.64	6.38	7–41	1.07
Release type
Standard supervision	0.67	0.47	0–1	0.67	0.47	0–1	0.68	0.47	0–1	−0.11
Community program	0.20	0.40	0–1	0.17	0.38	0–1	0.23	0.42	0–1	−1.50
ISR	0.07	0.26	0–1	0.12	0.32	0–1	0.03	0.17	0–1	3.31*
Discharge	0.06	0.23	0–1	0.05	0.21	0–1	0.07	0.25	0–1	−0.88
High school diploma	0.77	0.43	0–1	0.81	0.39	0–1	0.72	0.45	0–1	2.13*
Married	0.08	0.26	0–1	0.06	0.23	0–1	0.10	0.29	0–1	−1.52
Age in years	31.49	9.03	19–60	31.48	9.61	19–60	31.50	8.44	19–55	−0.02
Female	0.50	0.50	0–1	–	–	–	–	–	–	–
Race
White	0.49	0.50	0–1	0.39	0.49	0–1	0.59	0.49	0–1	−4.07*
Black	0.30	0.46	0–1	0.43	0.50	0–1	0.17	0.38	0–1	5.79*
Native American	0.17	0.37	0–1	0.12	0.33	0–1	0.22	0.41	0–1	−2.56*
Other race	0.05	0.21	0–1	0.07	0.25	0–1	0.03	0.16	0–1	1.93†

Note. LS/CMI = level of service/case management inventory; ISR = intensive supervised release.

p < .05.

†

p < .10.

Table 1 reveals several important gender differences in the variables used for the study. Women were significantly less likely than men to be rearrested (t = 4.04), reconvicted (t = 2.96), reincarcerated for a new felony (t = 2.90), or reincarcerated due to supervised release revocation (t = 2.43). Both genders were equally likely to participate in co-offenses (t = −0.41) and had similar average numbers of co-offenders (t = 0.08), but women had significantly lower roles in the offense (t = 3.83) and were significantly more likely to co-offend with romantic partners or family members (t = −3.37). Women were significantly less likely than men to have committed robbery offenses (t = 2.07), to have been committed from the Twin Cities metropolitan area (t = 2.93), be released on intensive supervised release (ISR, t = 3.31), or have a high school diploma (t = 2.13). Women had significantly shorter lengths of imprisonment than men (t = 3.65), had significantly fewer prior convictions (t = 3.79), and were significantly more likely to participate in treatment (t = −2.56). Finally, Table 1 shows that, unlike men, a majority of women were White or Native American, while Blacks and other non-White ethnicities were more highly represented among men.

Dependent Variables

Recidivism was measured as a rearrest for a new offense, which was obtained electronically from the Minnesota Bureau of Criminal Apprehension (BCA). The outcome included both “status” variables indicating whether an individual recidivated and “time” variables measuring the number of months between release and the first recidivism event. Recidivism data were collected through March 1, 2020, resulting in a follow-up period between 22 and 73 months, with an average follow-up period of 52 months. In supplemental analyses, additional measures of recidivism are examined: reconviction for a new offense (also obtained from BCA) and reincarceration either for a new felony offense or for a violation of supervised release (both obtained from COMS).

Independent Variables

Several aspects of co-offending within the initial burglary and/or robbery offense(s) are used as independent variables. The existence, identities, and actions of co-offenders were determined by examining the criminal complaint, the defendant’s description of the offense contained in the PSI report, and, in some cases, the co-offender’s description of the offense contained in his or her PSI report. First is a binary measure indicating whether any of the offenses for which the individual was incarcerated were committed with at least one other person. An offense was considered a co-offense if the description of the offense provided by either the police (in the criminal complaint) or the defendant (in the PSI report) mentioned the participation of others, even if those others were not convicted or even identified by police. Over half of the sample (61%) worked with co-offenders in at least one offense. The second dependent variable is a continuous variable measuring the average number of co-offenders involved in the offenses for which the subject was incarcerated. The subjects’ average number of co-offenders ranged from 0 to 6, with an average of 1.03.

Third is an ordinal variable that measures the highest role in any of the offenses for which the subject was incarcerated, based on their involvement in the planning and commission of the offense as described in the criminal complaint and/or the defendant’s or co-offender’s accounts of the offense in the PSI report. The variable includes three categories: (1) the subject was an accomplice to his or her co-offenders, in that they acted as a decoy, lookout, or lure while the co-offender(s) carried out a more active or violent role in the offense; (2) the subject had an equal role to his or her co-offenders; or (3) the subject had a major role as an instigator of the offense, in that they carried out the most active or violent part of the offense, either alone or with co-offenders acting as accomplices. Over half (60%) of the sample was either solo offenders or had a major role, while 27% had an equal role to co-offenders and 13% were accomplices.

The final independent variable measures whether the subject participated in an offense with a romantic partner or family member. The relationship between the study subject and each co-offender was captured; the dependent variable was coded as 1 if at least one co-offender was a romantic partner or family member and 0 if none of the co-offenders were romantic partners or family members or if there were no co-offenders. This information was typically found in the subject’s PSI, but was sometimes also available in the criminal complaint or the co-offender’s PSI. Twenty percent of the sample worked with romantic partners or family members in at least one of the offenses for which they were incarcerated.

Control Variables

To control for risk of recidivism, the analyses controlled for the most recent Level of Service/Case Management Inventory (LS/CMI) before release from prison. In addition, there were four continuous variables measuring (1) the length of the prison stay in months, (2) the number of discipline convictions during the current incarceration, (3) the number of visits received during the current incarceration, and (4) the number of previous felony convictions, not counting the current conviction(s). Next were six binary variables indicating whether the person was (1) incarcerated for robbery (compared to those incarcerated for burglary only), (2) incarcerated for a new commitment (compared to a release violation), (3) was committed from the 7-county Twin Cities metropolitan area, (4) had a high school diploma or GED before their release from prison, (5) was married, and (6) completed chemical dependency, sex offender, or educational programming before release from prison. Release type was measured as a series of binary variables indicating whether the person was released to (1) a community program, (2) intensive supervised release (ISR), (3) standard supervision (reference group), or (4) was discharged with no supervision.² Finally, the analyses controlled for race (White, Black, Native American, or another race, with White serving as the reference category), gender (female, compared to males), and age in years at the time of release.

Data Analysis

Because information on the timing of recidivism events was available, this study used survival analysis. Survival analyses are preferable over logistic regression because they allow for an examination of not only whether individuals commit a new offense, but also how quickly they do so. In particular, this study used Cox regression models, which employ both “status” and “time” variables. The “status” variable was a binary measure with a value of 1 if the event occurred. The “time” variable measured the amount of time (in months) between the release date and the date of the first recidivism event (or March 1, 2020, for those who did not recidivate). Checks for collinearity were conducted and no problems were found; all tolerance values were above 0.4.

Results

Bivariate Results

Table 2 displays the bivariate correlations between co-offending and recidivism. Among the total sample, rearrest and reconviction were negatively related to co-offending, the number of co-offenders, and participation in crime with romantic partners or other family members, and positively related to higher roles in the offense. Reincarceration for a new felony showed similar patterns, although only the correlation with number of co-offenders reached statistical significance. In contrast, supervised release revocation was not significantly related to any aspect of co-offending. When examining the gender-specific bivariate correlations, it appears that these relationships were more salient among females than among males. All aspects of co-offending were significantly or marginally significantly related to rearrest and reconviction among women, while fewer relationships showed statistical significance among men.

Table 2.

Bivariate Correlations between Co-Offending and Recidivism.

	Rearrest	Reconviction	Reincarceration	Revocation
Total sample
Co-offending	−0.116*	−0.124*	−0.085†	−0.006
Number of co-offenders	−0.106*	−0.103*	−0.136**	0.027
Highest role in an offense	0.200**	0.185**	0.098†	0.037
Any co-offense with romantic/family	−0.108*	−0.087†	−0.054	0.013
Males
Co-offending	−0.102	−0.086	−0.118†	−0.012
Number of co-offenders	−0.098	−0.081	−0.181*	−0.008
Highest role in an offense	0.164*	0.094	0.115	0.038
Any co-offense with romantic/family	0.022	0.115	0.064	0.111
Females
Co-offending	−0.125†	−0.157*	−0.041	0.006
Number of co-offenders	−0.121†	−0.131†	−0.080	0.065
Highest role in an offense	0.172*	0.213**	0.036	−0.005
Any co-offense with romantic/family	−0.141*	−0.197**	−0.123†	−0.027

p < .05. **p < .01.

†

p < .10.

Multivariate Results

The results of the Cox regression models predicting rearrest for the total sample are presented in Table 3. Both the presence of co-offenders (hazard ratio = 0.73, p = .015) and the number of co-offenders (hazard ratio = 0.85, p = .007) were related to lower risk of rearrest. Higher roles in the offense were related to higher risk of rearrest (hazard ratio = 1.37, p = .002). However, there was no relationship between rearrest and co-offending specifically with romantic partners or family members (hazard ratio = 0.83, p = .257).

Table 3.

Cox Regression Models Predicting Rearrest, Total Sample.

	Model 1			Model 2			Model 3			Model 4
	HR	SE	p	HR	SE	p	HR	SE	p	HR	SE	p
Any co-offense	0.73	0.13	.015*	–	–	–	–	–	–	–	–	–
Average number of co-offenders	–	–	–	0.85	0.06	.007**	–	–	–	–	–	–
Highest role in an offense	–	–	–	–	–	–	1.37	0.10	.002**	–	–	–
Any co-offense with romantic/family	–	–	–	–	–	–	–	–	–	0.83	0.17	.257
Robbery offense	1.00	0.17	.987	1.02	0.17	.923	0.98	0.17	.887	0.96	0.17	.812
New commitment	1.29	0.22	.243	1.26	0.22	.289	1.24	0.22	.311	1.25	0.22	.305
Metro commitment	1.15	0.14	.317	1.15	0.14	.334	1.14	0.14	.363	1.14	0.14	.372
Length of stay (months)	1.00	0.004	.368	1.00	0.004	.337	1.00	0.004	.179	1.00	0.004	.347
Discipline	1.00	0.01	.968	1.00	0.01	.878	1.00	0.01	.968	1.00	0.01	.999
Treatment	0.72	0.16	.040*	0.72	0.16	.043*	0.73	0.16	.055†	0.74	0.16	.066†
Visitation	1.00	0.002	.177	1.00	0.002	.238	1.00	0.002	.172	1.00	0.002	.265
Prior convictions	1.08	0.02	.000***	1.08	0.02	.000***	1.08	0.02	.000***	1.08	0.02	.000***
LS/CMI	1.04	0.02	.030*	1.01	0.01	.022*	1.03	0.02	.037*	1.04	0.01	.022*
Release type
Community program	0.93	0.17	.654	0.87	0.17	.397	0.91	0.17	.587	0.88	0.17	.465
ISR	0.99	0.25	.975	0.95	0.25	.840	1.00	0.24	.988	1.02	0.25	.941
Discharge	0.96	0.34	.897	0.93	0.35	.827	0.93	0.34	.836	1.02	0.35	.960
High school diploma	0.98	0.15	.900	0.96	0.15	.805	0.99	0.15	.957	0.99	0.15	.948
Married	0.64	0.26	.084†	0.62	0.26	.060†	0.63	0.26	.075†	0.66	0.26	.102
Age in years	0.97	0.01	.004**	0.98	0.01	.005**	0.97	0.01	.003**	0.98	0.01	.011*
Female	0.68	0.13	.004**	0.67	0.13	.003**	0.74	0.14	.029*	0.70	0.14	.009**
Race
Black	1.17	0.17	.359	1.14	0.17	.420	1.19	0.17	.311	1.16	0.17	.365
Native American	1.16	0.18	.417	1.18	0.18	.364	1.12	0.18	.516	1.19	0.18	.334
Other race	0.93	0.33	.819	0.92	0.33	.809	0.99	0.33	.982	0.92	0.33	.787

Note. LS/CMI = level of service/case management inventory; ISR = intensive supervised release.

p < .05. **p < .01. ***p < .001.

†

p < .10.

Several control variables were also related to rearrest among the total sample. Those who completed treatment (hazard ratio = 0.72, p = .040), were younger (hazard ratio = 0.97, p = .004), or were married (hazard ratio = 0.64, p = .084) had lower risk of rearrest, while higher LS/CMI scores (hazard ratio = 1.04, p = .030) and more prior convictions (hazard ratio = 1.08, p < .001) were related to higher risk of rearrest.

Next, gender-specific analyses were conducted to determine whether these relationships between co-offending and rearrest vary by gender. The results for the subsample of women are presented in Table 4. The results show similar relationships between rearrest and co-offending as were reported above. The presence of co-offenders (hazard ratio = 0.66, p = .040), the number of co-offenders (hazard ratio = 0.77, p = .012), and co-offending with romantic partners or family members (hazard ratio = 0.52, p = .004) were related to lower risk of rearrest, while higher roles in the offense were related to higher risk (hazard ratio = 1.53, p = .002). Two control variables were significantly related to rearrest among women; prior convictions were related to higher risk of rearrest (hazard ratio = 1.08, p = .019), while age was negatively related to risk of rearrest (hazard ratio = 0.96, p = .005).

Table 4.

Cox Regression Models Predicting Rearrest, Females.

	Model 1			Model 2			Model 3			Model 4
	HR	SE	p	HR	SE	p	HR	SE	p	HR	SE	p
Any co-offense	0.66	0.20	.040*	–	–	–	–	–	–	–	–	–
Average number of co-offenders	–	–	–	0.77	0.10	.012*	–	–	–	–	–	–
Highest role in an offense	–	–	–	–	–	–	1.53	0.13	.002**	–	–	–
Any co-offense with romantic/family	–	–	–	–	–	–	–	–	–	0.52	0.23	.004**
Robbery offense	0.93	0.25	.782	0.95	0.25	.831	1.00	0.26	.992	0.91	0.26	.717
New commitment	0.78	0.42	.540	0.76	0.42	.517	0.67	0.42	.344	0.77	0.42	.524
Metro commitment	0.92	0.23	.729	0.96	0.23	.847	0.91	0.23	.682	0.90	0.23	.661
Length of stay (months)	0.99	0.01	.364	0.99	0.01	.318	0.99	0.01	.227	0.99	0.01	.420
Discipline	1.00	0.01	.675	1.01	0.01	.508	1.00	0.01	.713	1.00	0.01	.903
Treatment	0.95	0.33	.872	0.97	0.33	.930	0.95	0.33	.875	1.05	0.33	.882
Visitation	1.00	0.003	.693	1.00	0.003	.487	1.00	0.003	.901	1.00	0.003	.918
Prior convictions	1.08	0.03	.019*	1.08	0.03	.022*	1.08	0.03	.016*	1.08	0.03	.021*
LS/CMI	1.02	0.02	.275	1.03	0.02	.161	1.02	0.02	.488	1.01	0.02	.491
Release type
Community program	0.88	0.25	.609	0.85	0.25	.515	0.91	0.25	.717	0.89	0.25	.648
ISR	1.92	0.56	.242	2.63	0.58	.097†	1.60	0.56	.397	1.96	0.56	.228
Discharge	0.87	0.56	.796	0.86	0.56	.794	0.72	0.56	.550	1.09	0.54	.880
High school diploma	0.80	0.22	.298	0.82	0.22	.363	0.79	0.22	.282	0.79	0.22	.279
Married	0.83	0.32	.563	0.79	0.32	.452	0.80	0.32	.475	0.86	0.32	.645
Age in years	0.96	0.02	.005**	0.96	0.02	.004**	0.96	0.02	.002**	0.95	0.02	.002**
Race
Black	0.89	0.31	.696	0.88	0.31	.668	0.85	0.32	.612	0.88	0.31	.679
Native American	1.02	0.24	.936	1.00	0.24	.999	0.95	0.24	.835	1.05	0.24	.829
Other race	0.42	0.77	.250	0.49	0.78	.350	0.44	0.76	.275	0.30	0.77	.114

LS/CMI = level of service/case management inventory; ISR = intensive supervised release.

p < .05. **p < .01. ***p < .001.

†

p < .10.

Finally, the results of the models predicting rearrest among men are displayed in Table 5. Unlike the results for the total sample and the subsample of women, none of the co-offending variables were related to risk of rearrest among men. In regards to control variables, treatment (hazard ratio = 0.67, p = .043), visitation (hazard ratio = 0.99, p = .067), and being married (hazard ratio = 0.41, p = .043) were related to lower risk of rearrest. Men with higher LS/CMI scores (hazard ratio = 1.05, p = .050) and more prior convictions (hazard ratio = 1.07, p = .002) had higher risk of rearrest.

Table 5.

Cox Regression Models Predicting Rearrest, Males.

	Model 1			Model 2			Model 3			Model 4
	HR	SE	p	HR	SE	p	HR	SE	p	HR	SE	p
Any co-offense	0.80	0.18	.207	–	–	–	–	–	–	–	–	–
Average number of co-offenders	–	–	–	0.90	0.08	.199	–	–	–	–	–	–
Highest role in an offense	–	–	–	–	–	–	1.21	0.15	.205	–	–	–
Any co-offense with romantic/family	–	–	–	–	–	–	–	–	–	1.46	0.26	.142
Robbery offense	0.99	0.23	.962	1.01	0.24	.957	0.94	0.23	.799	0.89	0.23	.610
New commitment	1.49	0.27	.137	1.46	0.27	.154	1.48	0.27	.143	1.53	0.27	.113
Metro commitment	1.38	0.20	.113	1.36	0.20	.133	1.39	0.20	.103	1.39	0.20	.102
Length of stay (months)	1.00	0.01	.694	1.00	0.01	.663	1.00	0.01	.563	1.00	0.01	.753
Discipline	1.00	0.01	.977	1.00	0.01	.999	1.00	0.01	.963	1.00	0.01	.900
Treatment	0.67	0.20	.043*	0.68	0.20	.048*	0.68	0.20	.051†	0.68	0.20	.051†
Visitation	0.99	0.002	.067†	1.00	0.002	.079†	1.00	0.002	.076†	1.00	0.002	.073†
Prior convictions	1.07	0.02	.002**	1.07	0.02	.002**	1.08	0.02	.001**	1.08	0.02	.001**
LS/CMI	1.05	0.02	.050†	1.05	0.02	.050†	1.05	0.02	.045*	1.06	0.02	.028*
Release type
Community program	0.92	0.24	0.74	0.86	0.24	.538	0.90	0.24	.658	0.86	0.24	.515
ISR	0.89	0.28	0.68	0.85	0.28	.551	0.92	0.28	.752	0.56	0.28	.589
Discharge	1.07	0.45	0.88	1.03	0.45	.954	1.09	0.45	.856	1.14	0.47	.788
High school diploma	1.14	0.22	.552	1.11	0.23	.641	1.17	0.22	.496	1.13	0.22	.587
Married	0.41	0.44	.043*	0.40	0.44	.038*	0.42	0.44	.045*	0.39	0.44	.032*
Age in years	0.99	0.01	.256	0.99	0.01	.268	1.00	0.01	.287	0.99	0.01	.394
Race
Black	1.33	0.22	0.19	1.32	0.22	.205	1.35	0.22	.172	1.32	0.22	.198
Native American	1.24	0.28	0.45	1.28	0.28	.377	1.25	0.28	.427	1.27	0.28	.406
Other race	1.26	0.39	0.55	1.25	0.39	.570	1.33	0.38	.452	1.41	0.39	.377

Note. LS/CMI = level of service/case management inventory; ISR = intensive supervised release.

p < .05. **p < .01. ***p < .001.

†

p < .10.

Supplemental Analyses

Alternative measures of recidivism

To test whether the results are robust to the measurement of recidivism, the relationships were examined when predicting reconviction, reincarceration for a new felony, and reincarceration for a supervised release violation. Table 6 summarizes the results found in those models (full models are available upon request). First, none of the co-offending variables were significantly related to return to prison for a supervised release violation, although the binary measure of co-offending approached significance among the full sample. Second, the results for reconviction were similar to those for arrest, with reconviction less likely when women co-offended, had more co-offenders, and co-offended with romantic partners or family members, and more likely when women had higher roles in the offense. However, men who co-offended with romantic partners or family members were more likely to be reconvicted.

Table 6.

Summary of Cox Regression Models Predicting Alternative Measures of Recidivism.

	Total	Females	Males
Reconviction
Co-offending	0.68 (0.15)**	0.55 (0.22)**	0.83 (0.21)
Number of co-offenders	0.84 (0.07)**	0.69 (0.12)**	0.94 (0.09)
Highest role in an offense	1.28 (0.07)***	1.35 (0.10)**	1.17 (0.10)
Any co-offense with romantic/family	0.81 (0.18)	0.44 (0.27)**	2.07 (0.27)**
Reincarceration for a new felony
Co-offending	0.64 (0.24)†	0.86 (0.38)	0.51 (0.33)*
Number of co-offenders	0.70 (0.12)**	0.82 (0.21)	0.59 (0.17)**
Highest role in an offense	1.33 (0.11)*	1.10 (0.16)	1.68 (0.18)**
Any co-offense with romantic/family	0.77 (0.30)	0.36 (0.51)*	1.65 (0.40)
Revocation of supervised release
Co-offending	0.72 (0.17)†	0.78 (0.26)	0.71 (0.24)
Number of co-offenders	0.95 (0.07)	1.01 (0.12)	1.05 (0.11)
Highest role in an offense	1.07 (0.08)	1.08 (0.11)	1.04 (0.11)
Any co-offense with romantic/family	0.88 (0.19)	0.67 (0.28)	1.09 (0.28)

Note. Hazard ratios are presented with standard errors in parentheses.

p < .05. **p < .01. ***p < .001.

†

p < .10.

Finally, the results for reincarceration for a new felony were markedly different than the results for rearrest and reconviction. Co-offending was more influential toward this outcome among men. Men who co-offended and who co-offended with larger groups were less likely to return to prison for a new felony sentence, while this was more likely when men had higher roles in the offense. Among women, reincarceration for a new felony was lower after co-offending with romantic partners or family members; the relationship to co-offenders was not related to reincarceration among men.

Alternative measurement of role in the offense

Because solo offenders are by default always in a leadership role, and may differ from those who lead criminal offenses with co-offenders, supplemental analyses (available upon request) were conducted in which those with no co-offenses were removed from the sample. When examining only those who worked with co-offenders, higher roles were related to greater risk of rearrest (hazard ratio = 1.328, p < .05) and reconviction (hazard ratio = 1.326, p < .05), but not reincarceration for a new felony (hazard ratio = 1.397, p > .10) or supervised release revocation (hazard ratio = 0.943, p > .10). As with the main analyses, the significant relationships between role in the offense and recidivism were only observed among women. The null finding regarding reincarceration for a new felony—which is different from the results presented in the main analyses examining the full sample—may indicate that more serious reoffending (which is more likely to result in a new prison sentence) is less influenced by one’s role in prior co-offenses than is less serious reoffending. On the other hand, the null finding could be due to the reduction in statistical power that comes with the use of a smaller sample, especially considering that reincarceration was rarer than rearrest or reconviction.

Discussion

This study examined whether people who commit crimes with co-offenders are more or less likely to reoffend, and whether that relationship is gendered. First, partially consistent with the expectations presented in Hypotheses 1 and 2, rearrest and reconviction were lower among those who co-offended and those with larger co-offending groups. Further, in line with Hypothesis 5, this relationship was only observed among women—the relationship was null when examining men. The gendered nature of the relationship between co-offending and rearrest informs the theoretical debate regarding whether co-offending increases recidivism. Women who co-offend may be drawn into these crimes by co-offenders when they are less inclined to commit crime alone, thereby having a lower likelihood of recidivating than women who commit solo offenses (McGloin & Piquero, 2009; Warr, 2002). Further, women who are drawn into crimes by co-offenders may not be particularly likely to be recruited for new criminal acts by those same individuals (see McGloin et al., 2008). On the other hand, men who co-offend may do so because of opportunity and may be equally likely to engage in solo offending, leading to the same recidivism risk as incarcerated males who committed their recent offenses alone. The results reinforce the argument posited by feminist criminologists that gender is vital for understanding multiple aspects of criminal behavior including recidivism and desistance and for designing treatment programs (e.g., Belknap, 2015; Chesney-Lind & Morash, 2013; Van Voorhis et al., 2010).

Partially consistent with Hypothesis 3, the results showed that—among women—the role in the offense was related to recidivism. Women who acted as accomplices were less likely to recidivate, while recidivism was more likely among women whose roles were more central to the offense. This finding is consistent with prior research on women’s involvement in crime, which suggests that many women are pulled into crime as accomplices and may not have been drawn to crime by their own individual motivations (e.g., Becker & McCorkel, 2011; McNeeley, 2019; Mullins & Wright, 2003). Importantly, women who are drawn into crime by their peers—who are likely to play secondary roles in the offenses—may be less motivated to engage in crimes more generally (McGloin & Piquero, 2009; Warr, 2002), making them more motivated to desist and making the desistance process easier. On the other hand, a man’s role in the offense was not related to recidivism, which suggests that working with criminal mentors may influence one’s own criminal motivation so that male accomplices and instigators have similar outcomes (see, for example, Morselli et al., 2006).

Finally, Hypothesis 4 predicted the relationship between the subject and his or her co-offenders would be related to future offending. This relationship was found to be significant, but not in the expected direction, and only among women. In particular, women who co-offended with romantic partners or family members were less likely to be rearrested than women who worked alone or with other acquaintances. It was hypothesized that these relationships would lead to reoffending; however, it is possible that the close, long-term nature of family and romantic relationships allows for greater recognition of the potential for corruption and therefore the need to cut ties with the person in order to be successful after release (e.g., Knight & West, 1975). Notably, this was only observed among women; this is in line with prior research on female desisters (Berg & Cobbina, 2017). This result speaks to the importance of romantic relationships for understanding women’s involvement in crime; women who co-offend with romantic partners may have been unlikely to engage in crime otherwise, and may have little motivation to engage in crime after the relationship ends.

However, the supplemental results regarding the relationship between co-offending and other measures of recidivism warrant further research. First, unlike rearrest and reconviction, co-offending was only related to reincarceration for a new felony among men. This indicates that co-offending may be related to not only reoffending but also the severity of the new offense. Some men who are involved in lower-level crime may be pulled into more serious crime by co-offenders. After release from prison, those men may continue engaging in lower-level crime but may be unlikely to recommit felonies once the ties with their co-offenders are broken. This dynamic may be similar to the process that leads to a negative relationship between co-offending and rearrest among women. Second, men were more likely to be reconvicted after a co-offense with a romantic partner or family member. This is consistent with life-course theories suggesting desistance among men is facilitated by relationships with prosocial romantic partners (e.g., Sampson & Laub, 1993). Additionally, co-offending had no relationship with supervised release revocation among either gender; future research should explore the influence of peers on one’s failure or success at following the technical conditions of supervised release.

Finally, the effects of some other correlates of recidivism varied based on gender. First, marital status was related to rearrest among men, but not among women. This is in line with prior research noting that the influence of marriage explained by Sampson and Laub (1993) applies more greatly to men than women (e.g., Giordano et al., 2002). Second, the relationship between visitation and rearrest was only significant among men, in line with a recent meta-analysis that found the effect of visitation on outcomes such as institutional misconduct and recidivism varies by gender (Mitchell et al., 2016). Third, men who completed chemical dependency, sex offender, and/or educational treatment were less likely to be rearrested, but this relationship was not found among women. This finding points to a need to develop and implement gender-responsive programming that addresses women’s unique risk factors (e.g., Van Voorhis et al., 2010).

Limitations and Suggestions for Future Research

While the results of this study provide important information on the relationship between co-offending and recidivism, there are limitations that must be acknowledged. First, due to the focus on burglary and robbery offenses, the results may not be generalizable to all incarcerated people, regardless of gender. Future research should replicate this study while examining other crime types. Second, the study relied on official measures of recidivism and only included recidivism that occurred in Minnesota. Therefore, some reoffending may not have been accounted for in the analyses. Notably, if those who co-offend are less likely to be arrested, as suggested by prior research (e.g., Bouchard & Nguyen, 2010; Lantz, 2020a; Morselli et al., 2006), this may partially explain the negative relationship between co-offending and recidivism found here. Future research should explore using alternative measures, such as self-reported data from releasees, to better capture all reoffending. Third, much of the research on co-offending examines juveniles. This study focused on co-offending and recidivism among adults, which could account for the unexpected results. Future research should examine whether the relationship between juvenile co-offending and later criminal behavior is gendered.

Fourth, this study focused on individuals who had been incarcerated for the offense in question; therefore, it is possible that the sample represents individuals who are more central within crime groups and/or those who are more likely to instigate criminal offenses, as they may be more likely than their subordinates to face a prison sentence. Indeed, only 13% of the sample acted as accomplices in all the offenses included in the study. Therefore, the results may not be completely generalizable to individuals who are more subordinate within a criminal network. Fifth, because the study only examines co-offending in the present offense(s) for which a person was incarcerated—not their entire criminal history—it is possible that some who are identified as solo offenders in this study have co-offended in the past, or vice versa. Therefore, the precise nature of the relationship between co-offending and reoffending could not be fully measured in this study.

Relatedly, past work has shown the importance of understanding aspects of co-offending networks—such as the group size, strength of ties, and breakdown of gender and age—for understanding how co-offending impacts outcomes (e.g., Andersen, 2019; Lantz, 2020b: Lantz & Hutchison, 2015; Ouellet et al., 2013). For example, Andersen (2019) confirmed that recidivism was more likely for those had more direct co-offending ties (i.e., connections with more unique co-offenders), even though the study did not observe repeated co-offenses with the co-offenders acquired before incarceration. Unfortunately, this study was not able to construct networks between the study subjects and the co-offenders they worked with across their criminal careers. Future research should continue to explore how the gendered nature of co-offending networks might differentially impact men’s and women’s later criminal behavior.

Conclusion

Despite these limitations, this study provides important insights regarding the relationship between co-offending and recidivism. The results suggested several aspects of co-offending—the number of co-offenders, the relationship to the co-offender(s), and the role in the offense—may shape future offending. Importantly, these relationships were more consistently observed for women than for men, highlighting the importance of considering gender when attempting to better understand risk for recidivism. The results suggest the presence or absence of co-offenders could be an indicator to use when assessing risk among women. Additionally, if future research confirms that women who commit offenses alone are more likely to recidivate, it might be beneficial for correctional programs designed for women to focus on women’s solo offending.

At the same time, sometimes women known to have co-offended will be prioritized for treatment (for example, because they are known to have high risk of recidivism). In these cases, it may be beneficial for these women to participate in programs that address the social aspects that can lead to co-offending. Association with anti-social peers has been identified as one of the most important risk factors for recidivism (Andrews et al., 2006). However, there are few correctional programs that target this criminogenic need. Those who commit crimes with others may be especially in need of treatment that addresses how to navigate relationships with others in prosocial ways. This study suggests programs designed to address anti-social peer influence should be gender-responsive and should focus on the dynamics between women offenders and their romantic partners and family members that may lead to co-offending.

Footnotes

Declaration of Conflicting Interests

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

Funding

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

ORCID iD

Susan McNeeley

Notes

Author Biography

Susan McNeeley is a senior research analyst with the Minnesota Department of Corrections. In addition to corrections, her research interests include criminological theory and victimology. Her recent work has been published in Justice Quarterly, Journal of Interpersonal Violence, Journal of Criminal Justice, and Criminal Justice and Behavior.

References

Akers

R. L.

(1998). Social learning and social structure: A general theory of crime and deviance. Northeastern University Press.

Alarid

L. F.

Marquart

J. W.

Burton

V. S.

Cullen

F. T.

Cuvelier

(1996). Women’s roles in serious offenses: A study of adult felons. Justice Quarterly, 13, 431–454.

Andersen

S. A.

(2019). Partners in crime? Post-release recidivism among solo and co-offenders in Norway. Nordic Journal of Criminology, 20, 112–137.

Andrews

D. A.

Bonta

Wormith

J. S.

(2006). The recent past and near future of risk and/or need assessment. Crime & Delinquency, 52, 7–27.

Becker

McCorkel

J. A.

(2011). The gender of criminal opportunity: The impact of male co-offenders on women’s crime. Feminist Criminology, 6, 79–110.

Belknap

(2015). The invisible woman: Gender, crime, and justice (4th ed.). Cengage Learning.

Berg

M. T.

Cobbina

J. E.

(2017). Cognitive transformation, social ecological settings, and the reentry outcomes of women offenders. Crime & Delinquency, 63, 1522–1546.

Bronson

Carson

E. A.

(2019). Prisoners in 2017. U.S. Department of Justice, Office of Justice Programs.

Bouchard

Nguyen

(2010). Is it who you know, or how many that counts? Criminal networks and cost avoidance in a sample of young offenders. Justice Quarterly, 27, 130–158.

10.

Carrington

P. J.

(2002). Group crime in Canada. Canadian Journal of Criminology, 44(3), 277–315.

11.

Carrington

P. J.

(2009). Co-offending and the development of the delinquent career. Criminology, 47, 1295–1329.

12.

Chesney-Lind

Morash

(2013). Transformative feminist criminology: A critical re-thinking of a discipline. Critical Criminology, 21, 287–304.

13.

Conway

K. P.

McCord

(2002). A longitudinal examination of the relation between co-offending with violent accomplices and violent crime. Aggressive Behavior, 28(2), 97–108.

14.

Cornish

D. B.

Clarke

R. V.

(1986). The reasoning criminal: Rational choice perspectives on offending. Transaction Publishers.

15.

Crosnoe

Erickson

K. G.

Dornbusch

S. M.

(2002). Protective functions of family relationships and school factors on the deviant behavior of adolescent boys and girls: Reducing the impact of risky friendships. Youth & Society, 33, 515–544.

16.

Giordano

P. C.

Cernkovich

S. A.

Rudolph

J. L.

(2002). Gender, crime, and desistance: Toward a theory of cognitive transformation. American Journal of Sociology, 107, 990–1064.

17.

Glueck

(1950). Unraveling Juvenile delinquency. Commonwealth Fund.

18.

Hindelang

M. J.

(1976). With a little help from their friends: Group participation in reported delinquency. British Journal of Criminology, 16, 109–125.

19.

Knight

B. J.

West

D. J.

(1975). Temporary and continuing delinquency. British Journal of Criminology, 15, 43–50.

20.

Koons-Witt

B. A.

Schram

P. J.

(2003). The prevalence and nature of violent offending by females. Journal of Criminal Justice, 31, 361–371.

21.

Kornhauser

R. R.

(1978). Social sources of delinquency: An appraisal of analytic models. University of Chicago Press.

22.

Lantz

(2020a). Co-offending and arrest: An examination of the “group hazard” hypothesis. Crime & Delinquency, 66, 572–600.

23.

Lantz

(2020b). Co-offending group composition and violence: The impact of sex, age, and group size in co-offending violence. Crime & Delinquency, 66(1), 93–122. https://doi.org/10.1177/0011128719834564

24.

Lantz

Hutchison

(2015). Co-offender ties and the criminal career: The relationship between co-offender group structure and the individual offender. Journal of Research in Crime and Delinquency, 52, 658–690.

25.

Lin

(2001). Social capital: A theory of social structure and action. Cambridge University Press.

26.

McCarthy

Hagan

(1995). Getting into street crime: The structure and process of criminal embeddedness. Social Science Research, 21, 63–95.

27.

McCarthy

Hagan

(2001). When crime pays: Capital, competence, and criminal success. Social Forces, 793, 1035–1059.

28.

McCord

Conway

K. P.

(2002). Patterns of juvenile delinquency and co-offending. In Waring

Weisburd

(Eds.), Advances in criminology theory, vol. 10: Crime and social organization (pp. 15–30). Transaction Publishers.

29.

McCord

Conway

K. P.

(2005). Co-offending and patterns of juvenile crime. National Institute of Justice.

30.

McGloin

J. M.

Nguyen

(2012). It was my idea: Considering the instigation of co-offending. Criminology, 50, 463–494.

31.

McGloin

J. M.

Piquero

A. M.

(2009). ‘I wasn’t alone’: Collective behavior and violent delinquency. Australian & New Zealand Journal of Criminology, 42, 336–353.

32.

McGloin

J. M.

Sullivan

C. J.

Piquero

A. R.

Bacon

(2008). Investigating the stability of co-offending and co-offenders among a sample of youthful offenders. Criminology, 46(1), 155–188.

33.

McNeeley

(2019). Gendered pathways into co-offending among a sample of adult burglary and robbery offenders. Crime & Delinquency, 65, 1711–1739.

34.

McNeeley

Warner

J. J.

(2015). Replication in criminology: A necessary practice. European Journal of Criminology, 12(5), 581–597.

35.

Miller

(1998). Up it up: Gender and the accomplishment of street robbery. Criminology, 36, 37–66.

36.

Mitchell

M. M.

Spooner

Jia

Zhang

(2016). The effect of prison visitation on reentry success: A meta-analysis. Journal of Criminal Justice, 47, 74–83.

37.

Morselli

Tremblay

McCarthy

(2006). Mentors and criminal achievement. Criminology, 44, 17–43.

38.

Mullins

C. W.

Wright

(2003). Gender, social networks, and residential burglary. Criminology, 41, 813–840.

39.

Muskens

Bogaerts

van Casteren

Labrijn

(2011). Adult female sexual offending: A comparison between co-offenders and solo offenders in a Dutch sample. Journal of Sexual Aggression, 17, 48–60.

40.

Ouellet

Boivin

Leclerc

Morselli

(2013). Friends with(out) benefits: Co-offending and re-arrest. Global Crime, 14, 141–154.

41.

Park

Morash

Stevens

(2010). Gender differences in predictors of assaultive behavior in late adolescence. Youth Violence and Juvenile Justice, 8(4), 314–331.

42.

Pattavina

Carkin

D. M.

Tracy

P. E.

(2017). Assessing the representativeness of NIBRS arrest data. Crime & Delinquency, 63(12), 1626–1652.

43.

Piquero

A. R.

Farrington

D. P.

Blumstein

(2007). Key issues in criminal career research. Cambridge University Press.

44.

Reiss

A. J.

(1986). Co-offending and criminal careers. In Morris

Tonry

M. H.

(Eds.), Crime and justice (Vol. 10, pp. 117–170). University of Chicago Press.

45.

Reiss

A. J.

Farrington

D. P.

(1991). Advancing knowledge about co-offending: Results from a prospective longitudinal survey of London males. Journal of Criminal Law and Criminology, 82, 360–395.

46.

Rennison

C. M.

(2009). A new look at the gender gap in offending. Women & Criminal Justice, 19(3), 171–190.

47.

Sampson

R. J.

Laub

(1993). Crime in the making: Pathways and turning points through life. Harvard University Press.

48.

Schwartz

Conover-Williams

Clemons

(2015). Thirty years of sex stratification in violent crime partnerships and groups. Feminist Criminology, 10, 60–91.

49.

Schwartz

Steffensmeier

(2017). Gendered opportunities and risk preferences for offending across the life course. Journal of Developmental and Life Course Criminology, 3, 126–150.

50.

Steffensmeier

Schwartz

Roche

(2013). Gender and 21st century corporate crime: Female involvement and gender gap in Enron-era corporate frauds. American Sociological Review, 78, 448–476.

51.

Sutherland

E. H.

(1947). Principles of criminology (4th ed.). Lippincott.

52.

van Mastrigt

S. B

. (2008). Co-offending: Relationships with age, gender, and crime type (PhD dissertation). University of Cambridge, UK.

53.

van Mastrigt

S. B.

Farrington

D. P.

(2009). Co-offending, age, gender, and crime type: Implications for criminal justice policy. British Journal of Criminology, 49, 552–573.

54.

Van Voorhis

Wright

E. M.

Salisbury

Bauman

(2010). Women’s risk factors and their contributions to existing risk/needs assessment: The current status of a gender-responsive supplement. Criminal Justice and Behavior, 37, 261–288.

55.

Wall

Power

Arbona

(1993). Susceptibility to anti-social peer pressure and its relation to acculturation in Mexican-American adolescents. Journal of Adolescent Research, 8, 403–448.

56.

Warr

(1998). Life-course transitions and desistance from crime. Criminology, 36, 183–216.

57.

Warr

(2002). Companions in crime: The social aspects of criminal conduct. Cambridge University Press.