The Gendered Nature of Criminogenic Thinking Patterns Among Justice-Involved Clients: A Pilot Study

Abstract

Because much of our understanding of criminogenic thinking (antisocial cognitions) has been based on male justice populations, questions remain about the applicability of this construct to justice-involved women. Based on an item-level analysis of 216 justice-involved clients, results of this pilot study suggest that criminogenic thinking in women is relevant, and both overlaps with and diverges from that of men. In fact, the predictive accuracy for rearrest attained with a gender-responsive model developed for women exceeded that of the corresponding model developed for men (area under the curve [AUC] = .86 vs. AUC = .67). We recommend the creation of parsimonious criminogenic thinking instruments that optimize predictive criterion validity. Gender-responsive scales that capture the gender-specificity that exists in criminogenic thinking patterns can assist in (a) optimizing the prediction of reoffending and (b) identifying essential constellations of treatment targets among forensic populations.

Keywords

criminal thinking risk assessment gender gender differences women offenders predictive validity

The principles articulated in the Risk–Need–Responsivity (RNR) model now guide much of the assessment and rehabilitation work conducted with forensic clients around the world. Despite their inception over 30 years ago, these core principles, described in detail elsewhere (Andrews et al., 1990, 2006; Bonta & Andrews, 2017), have retained their value in guiding effective correctional practice. Notwithstanding the longevity and robustness of RNR, pioneers and key proponents of the model like Steve Wormith were ultimately innovators, guided by empirical data, open to advancements in theory and practice, and embraced the pursuit of theoretical refinements as a productive and necessary part of scientific progress. Both our personal conversations with Steve and evidence from his own body of work support his quest for innovation.¹

Regrettably, some alternative theoretical frameworks emerging in the field of corrections—namely, related to gender-specificity and strength-based perspectives—have often been positioned as diametrically opposed to RNR. We would argue that this dissonance is largely unnecessary and instead we advocate for the idea of theoretical integration. In the spirit of innovation and complementarity, the current pilot study represents a potential extension of one element of the RNR model. Specifically, this study explores gender differences in the relationship between criminogenic thinking (often termed procriminal attitudes or antisocial cognitions) and recidivism in a sample of women and men under community supervision. We provide evidence to suggest that criminogenic thinking is relevant to the criminal conduct of both women and men as a global construct—at least as it is currently conceptualized in the literature. However, the systematic unpacking of the construct, coupled with the disaggregation of assessment data by gender, suggests considerable gender-specificity of criminogenic thinking patterns and their relationship to recidivism.

Criminogenic Thinking and RNr

Defined as cognitive patterns that facilitate antisocial, criminal, and self-destructive behavior, criminogenic thinking as a risk factor for recidivism derived initial empirical support from longitudinal studies with adult probationers (Andrews et al., 1985; Andrews & Wormith, 1984) and several meta-analytic reviews (Gendreau et al., 1996, 1997; L. Simourd & Andrews, 1994). Initially, the major risk factors for criminal conduct articulated under the RNR framework were grouped into two tiers, with criminogenic thinking (along with history of antisocial behavior, antisocial associates, and antisocial personality) considered to be among the “Big Four” risk factors most strongly linked with reoffending (Bonta & Andrews, 2017). The tier-level distinction has fallen out of favor to some extent, as the relative strength of these factors appears to vary across different forensic samples (Mitchell et al., 2016).

Regardless, criminogenic thinking remains a robust predictor of criminal behavior (Bonta & Andrews, 2017; Mills et al., 2004; D. J. Simourd & Olver, 2002; Walters, 2004) and institutional misconduct (Walters & Mandell, 2007; Walters & Schlauch, 2008), and is proposed as a key dynamic treatment target or criminogenic need in forensically oriented cognitive-behavioral therapy (CBT; Morgan et al., 2018; Tafrate et al., 2018). Criminogenic thinking patterns have also been associated with treatment attrition, suggesting that addressing such patterns early in the intervention process is apt to enhance treatment responsivity (Mitchell et al., 2013; Olver et al., 2011; Taxman et al., 2011).

Criminogenic Thinking in Practice: A Complex Landscape

Prior to the inception of the RNR principles, the construct of criminogenic thinking also emerged in the context of the cognitive revolution that occurred in psychotherapy beginning in the 1950s, while Ellis (1957, 1962) and A. T. Beck (1963, 1967) were formulating CBT as a treatment model for anxiety and depression. CBT was proposed as an alternative to strict behaviorism and psychoanalysis, the latter two being the dominant therapeutic approaches at the time. In a parallel yet distinct stream of literature, Sykes and Matza (1957), in one of the seminal works in this area, proposed that forensic clientele “neutralize” the wrongfulness of their criminal actions by embracing a unique constellation of cognitions that free them of guilt, self-blame, and disapproval from others.

Although restructuring dysfunctional thinking patterns is central to CBT-oriented interventions, a meaningful distinction can be made regarding the nature of the thinking patterns that should be targeted when addressing antisocial behavior versus common mental health disorders. Specifically, typical anxious and depressive thinking patterns overestimate and exaggerate potential dangers, emphasize self-blame, and undercut self-efficacy (Barlow et al., 2017; J. S. Beck, 2011; DiGiuseppe et al., 2014). In contrast, criminogenic thinking patterns involve a tendency to underestimate risk of difficulty, shift blame to others, and are unrealistically self-confident (Mitchell et al., 2015). Thus, an exclusive focus on mental health symptoms is ill-advised when working with forensic clients if a goal of treatment is to reduce their justice involvement. Even in clients who present with both mental health and criminal problems, the cognitions that amplify their dysphoric symptoms will be different from those that drive their criminal decisions.

Criminogenic thinking is multifaceted and is not adequately represented by a single cognitive pattern. However, in many forensic contexts, criminogenic thinking is measured on broad rater-based criminal risk/need assessment instruments such as the Level of Service/Case Management Inventory (LS/CMI; Andrews et al., 2004) family of tools and the Ohio Risk Assessment System (ORAS; Latessa et al., 2009). The necessity for these general risk assessment tools to assess multiple risk domains leaves limited opportunity for capturing an array of criminogenic thinking patterns in the space allotted to this risk factor on any given instrument. Consequently, broad criminal risk/need instruments capture criminogenic thinking as a global construct, represented by a small selection of items aggregated (often on the basis of face validity) into a single domain.

In contrast, a valuable aid in conceptualizing the breadth of thinking patterns hypothesized to foster criminal behavior can be found in the empirical literature that has emerged since the mid-1990s around a collection of self-report assessment instruments for use with justice-involved populations.² Because each of these instruments features multiple subscales, a sizable array of distinct criminogenic thinking patterns has emerged, including but not limited to identification with antisocial peers, justifying, excitement seeking, disregard for others, and hostility toward law and order. Several overarching cognitive themes have also been identified through factor analytic research. Walters (2019) has proposed two overlapping higher order thinking styles: Proactive (goal-directed thinking that reflects organization and planning) and Reactive (emotion-driven thinking that reflects impulsiveness, irresponsibility, and low self-control). In turn, Mandracchia and colleagues (2007) have proposed a tripartite model that includes themes of Cognitive Immaturity (a tendency for impulsivity and overgeneralization), Egocentrism (self-centeredness across situations and interactions), and Control (desire for power over others and the environment). For a detailed overview of the criminogenic thinking assessment literature, see Kroner and Morgan (2014).

Is the Criminogenic Thinking Construct Relevant to Girls and Women?

Over the past few decades, a notable controversy has emerged regarding the relevance of criminogenic thinking to justice-involved girls and women (O’Hagan et al., 2019). Largely based on qualitative accounts, some theorists suggest that such thinking patterns are not particularly germane to the criminal conduct of women, arguing that women become enmeshed in the justice system due to a unique constellation of systemic factors linked to patriarchal oppression: sexual victimization, intimate partner abuse, financial dependence, child care responsibilities, a desire to maintain relationships, and, ultimately, survival needs (e.g., Chesney-Lind & Shelden, 2014; Daly, 1992; Jones, 2011; Salisbury & Van Voorhis, 2009; Steffensmeier & Allan, 1996). Thus, many feminist scholars argue that an unfortunate by-product of focusing on criminogenic thinking in correctional assessment and treatment is the placement of unnecessary blame and responsibility on girls and women by decontextualizing their criminal behavior (Bloom et al., 2003; Hannah-Moffat, 2006; Van Dieten & King, 2014).

Further fueling the argument that criminogenic thinking is an androcentric construct is the fact that men are more highly represented in the justice system worldwide, and resultantly, criminogenic thinking models, formulations, and instruments have primarily been developed on male samples—potentially rendering these tools less relevant to women. Indeed, there is some evidence to suggest that gender-responsive risk assessment tools and correctional programming designed specifically for girls and women produce superior correctional outcomes to gender-neutral or inherently male-based approaches (Gobeil et al., 2016; Jones, 2011; Van Dieten & King, 2014; Van Voorhis et al., 2010).

In contrast, other scholars have long contended that the central predictors of criminal conduct, including antisocial cognitions, hold universally irrespective of gender (e.g., Bonta & Andrews, 2017). Meta-analytic evidence appears to support this position, with broadly defined, rater-based criminogenic thinking scores predicting general delinquency and recidivism outcomes equally well for girls/women and boys/men (Dowden & Andrews, 1999; Hubbard & Pratt, 2002; Olver et al., 2014; L. Simourd & Andrews, 1994). Nonetheless, the literature remains unclear about which elements of this global and sometimes inconsistently defined construct are more salient to one gender versus the other in the prediction of criminal behavior (O’Hagan et al., 2019).

Specialized self-report instruments can potentially provide opportunities to explore gender differences in overall criminogenic thinking and in specific thinking patterns. Regrettably, most evidence of gender differences in both prevalence and predictive validity is highly mixed and inconsistent across studies (e.g., Folk et al., 2018; Henning et al., 2009; Mitchell & Tafrate, 2012; O’Hagan et al., 2019; Tangney et al., 2012; Taxman et al., 2011; Walters & Lowenkamp, 2016). Possible reasons for these inconsistencies may be sample differences in age (adult vs. juvenile), offense type (drug-involved vs. violent vs. general), and setting (community vs. custody). However, a more likely factor is divergent instrument and subscale content: Although there is a degree of conceptual overlap in the patterns measured across criminogenic thinking instruments and subscales, not all instruments measure the same patterns. Gender differences in pathways to the criminal justice system may increase the salience of a pattern captured by a subscale on one instrument that is not captured on a different instrument. The challenge of drawing conclusions posed by sample, instrument, and subscale content differences is exacerbated by the paucity of studies exploring gender differences on any single instrument.

Despite the largely inconsistent findings described above, there is some preliminary quantitative evidence to suggest that attitudes overtly supportive of deviance and antisociality are more potent predictors of criminality among men/boys than women/girls (e.g., Heimer, 1996; O’Hagan et al., 2019). This finding is further supported by qualitative interview data with justice-involved girls, indicating that many do not have a profound attachment to their delinquent activities and tend to voice resentment about being labeled “delinquent” (Chesney-Lind & Shelden, 2014). A multidimensional scaling analysis by Jones and colleagues (2014) revealed that the theme of antisociality, largely driven by more overt forms of criminogenic thinking and violent offending, was only relevant to about half the justice-involved girls in their sample. The other half adhered to a gendered pathway into crime characterized by early abuse, poverty, minor forms of criminal behavior, and, importantly, the absence of antisocial cognitions as measured by a broad-based risk/need assessment tool.

Borrowing from the gender-responsive literature, it is plausible that instead of criminogenic thinking being nonexistent among a large proportion of women, that current conceptualizations of the construct (particularly those captured on broad-based measures) are omitting key elements that may be salient or specific to the criminal behavior of girls and women. These may include thinking patterns couched in survival motives, maladaptive attachment, or poor coping (e.g., justifying the commission of crime as a means of survival, to protect a loved one, or to preserve a significant relationship; Belknap, 2015; Bloom et al., 2003; O’Hagan et al., 2019).

A Brief Note About Instrument Construction and Overview of the Present Research

In the forensic area, two main methodological traditions in instrument development have emerged. The self-report criminogenic thinking instruments noted earlier generally adhere to a classical test construction methodology that prioritizes construct validity and adopts a factor analytic approach for identifying and confirming the measurement of latent constructs (i.e., underlying criminogenic thinking patterns) that exist in justice-involved populations. Once reliably established, instruments and subscales are primarily used to identify relevant treatment targets but have also been used to predict key outcomes such as treatment adherence and rearrest.

In contrast, risk assessment instrument development has prioritized criterion validity, the primary purpose traditionally being to predict an outcome of interest (e.g., recidivism) with greatest possible accuracy. With this approach, there is no particular expectation of measuring an underlying construct, and the goal of optimizing predictive validity trumps internal consistency and factor structure. Both predictive accuracy and efficiency of administration are maximized by including the smallest number of items measuring distinct constructs—goals that will deliberately decrease internal consistency.

Both traditions provide something of value to researchers and practitioners. In many settings, sorting the population of interest with the purpose of identifying individuals most likely to commit further offenses is paramount (criterion-referenced approach). In other environments, identifying specific treatment targets to guide case management and intervention is the ultimate aim of assessment (latent construct approach). In many contexts, a combination of the two goals is necessary and as such, a blending of the traditions is advisable.

In the present study, we begin by exploring gender differences on the Criminogenic Thinking Profile (CTP; Mitchell & Tafrate, 2012) from both a prevalence and prediction perspective (overall and at the subscale level). Next, through an analysis of individual criminogenic thinking items featured on the original CTP, and their capacity to predict rearrest, we deliberately employ a purely criterion-referenced instrument construction methodology to optimize prediction by gender and create new gender-responsive and gender-specific scales. We conclude our analyses by comparing the predictive utility of the gender-responsive and gender-specific scales with that of the original CTP and gender-neutral models.

Method

Participants and Setting

Participants were 216 consecutive admissions to a criminal justice day reporting center (DRC) in Connecticut that provides court-mandated case management, drug testing, community service opportunities, and psychoeducational groups focused on substance use, employment, and decision making. The DRC serves primarily adult female and male justice-involved clients arrested or charged with violent, drug, or property offenses (individuals arrested for the commission of a sexual offense are excluded from services). Clients may be serving probation, awaiting trial, or on parole. Each client is assigned a case manager, who coordinates services with the court and prepares reports on progress and compliance. Clients undergo a Level of Service Inventory–Revised (LSI-R; Andrews & Bonta, 1995) interview and complete the CTP (Mitchell & Tafrate, 2012) as part of standard intake procedures.

Female participants (n = 48) were a mean age of 22.83 years (SD = 10.07) at the time of data collection. In terms of ethnic background, 31.3% (n = 15) identified themselves as Latina, 47.9% (n = 23) as White, 14.6% (n = 7) as African American, and 6.3% (n = 3) as “Other”. Their mean educational level was 10.60 years (SD = 1.47), and 25% (n = 12) were working on a full- or part-time basis. With respect to criminal history, the mean number of arrests was 2.08 (SD = 6.28) and 20.2% (n = 10) had been referred to the DRC on a felony charge, with the remainder referred for a misdemeanor. According to the LSI-R administered at intake, 35.4% of women were classified as low risk, 14.6% as medium risk, and 31.3% as high risk (M = 22.79, SD = 9.76).³

Male participants (n = 168) were an average age of 24.72 years (SD = 9.54), with 44.6% (n = 75) self-identifying as Latino, 32.7% (n = 55) as White, 17.3% (n = 29) as African American, and 5.4% (n = 9) as “Other”. Their mean educational level was 10.77 years (SD = 1.32), and 26.8% (n = 48) were working on a full- or part-time basis. With respect to criminal history, the mean number of arrests was 4.08 (SD = 5.63), and 59.5% (n = 100) had been referred to the DRC on a felony charge. Based on intake LSI-R scores, 20.2% of male participants were classified as low risk, 31.0% as medium risk, and 45.8% as high risk (M = 26.87, SD = 7.17).

Significant gender differences emerged on three of the variables reviewed above. Compared with their female counterparts, male participants had more arrests, t(214) = 2.12, p = .036, d = 0.33; a higher mean risk score on the LSI-R, t(200) = 2.96, p = .003, d = 0.52; and a greater percentage had been referred to the DRC on a felony, χ² (1, 215) = 15.23, p < .001, Cramer’s V = .27.

The CTP

The CTP (Mitchell & Tafrate, 2012) is a 62-item self-report criminogenic thinking instrument that yields a total score as well as scores on eight thinking pattern subscales: (a) Disregard for Others (lack of concern and remorse), (b) Demand for Excitement (risk-taking, low tolerance for boredom), (c) Poor Judgment (underestimating negative outcomes from risky behaviors), (d) Emotionally Disengaged (lack of trust, avoidance of intimacy), (e) Parasitic/Exploitive (seeking to exploit people/situations), (f) Justifying (minimization of antisocial/self-destructive behavior), (g) Inability to Cope (giving up in the face of adversity), and (h) Grandiosity (inflating skills and personal qualities). Examples of CTP items are provided in Tables 1 to 3.

Table 1:

Gender-Neutral Scales and Items and Associated Predictive Validity by Gender

Subscales/items	Women (n = 48)		Men (n = 168)
Subscales/items	Rho	OR [95% CI]	Rho	OR [95% CI]
CTP subscales
Justifying	.28*	1.35 [1.00, 1.82]	.25***	1.20 [1.08, 1.35]
Poor Judgment	.34**	1.13 [0.95, 1.36]	.13*	1.08 [0.99, 1.18]
Items (n = 7)
It doesn’t make sense to plan for the future because you may not be around to see it	.23*	2.79 [1.08, 7.22]	.21**	1.77 [1.23, 2.54]
I won’t get caught again because I’m smarter than the police	.30*	5.00 [1.01, 24.77]	.12^†	1.51 [0.90, 2.53]
One of the problems I had in school is that I’m smarter than most of the teachers^a	.35**	4.45 [1.05, 18.86]	.16*	1.74 [1.01, 2.97]
Treatment programs can’t help me because I’m smarter than most of the counselors	.31*	2.25 [0.72, 6.97]	.21**	2.30 [1.31, 4.06]
It’s OK to break the law if you need to support your family	.19^†	1.93 [0.76, 4.94]	.21**	1.80 [1.19, 2.72]
It’s OK to use drugs sometimes because even the President has gotten high	.21^†	1.89 [0.67, 5.28]	.13*	1.64 [1.05, 2.56]
It’s OK to take things from people who have more than you	.30*	4.74 [1.11, 20.25]	.11^†	1.63 [0.91, 2.92]

Note. rho = 0.10 (small effect size), 0.30 (medium effect size), and (0.50 large effect size). Although items reaching p < .10 do not attain the traditionally accepted threshold for statistical significance, they were retained as per the liberal item selection criterion discussed in the Results section. OR = odds ratio; CI = confidence interval; CTP = Criminogenic Thinking Profile.

Item dropped from gender-responsive model and gender-neutral model for male sample due to multicollinearity.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Table 2:

Female-Specific Scales and Items and Associated Predictive Validity

Subscale/items	Rho	OR [95% CI]
CTP subscale
Disregard for Others	.27*	1.13 [0.99, 1.30]
Items (n = 12)
Getting drunk or high helps me deal with my problems	.22^†	1.62 [0.71, 3.71]
I cannot do boring things like chores around the house without being buzzed or high	.24*	2.26 [0.84, 6.08]
I don’t need an education because I will always find ways to make money	.35**	2.36 [0.87, 6.37]
The types of crimes that I commit won’t hurt me down the road	.24*	1.66 [0.80, 3.43]
Prisons and halfway houses are a good deal because they pay for your housing, meals, and medical expenses	.25*	2.29 [0.67, 7.81]
It doesn’t make sense to work full-time if you can get on a government program^a	.25*	2.29 [0.67, 7.81]
If you don’t lock your door, you deserve to get robbed	.38**	4.87 [1.14, 20.85]
I have more positive qualities than most people	−.20^†	0.46 [0.14, 1.49]
I am destined for greatness	−.30**	0.43 [0.18, 1.04]
Stealing from businesses doesn’t hurt anybody because insurance pays for it^a	.28*	4.34 [1.05, 17.83]
Case workers, probation officers, and counselors will never be able to help me	.23^†	2.34 [0.65, 8.44]
If someone wrongs me, I will get them back twice as bad	.33**	2.43 [0.94, 6.28]

Items eliminated from gender-responsive model for women due to multicollinearity with other items.

†

p < .10. *p < .05. **p < .01.

Table 3:

Male-Specific Scales and Items and Associated Predictive Validity

Subscale/items	Rho	OR [95% CI]
CTP subscale
Demand for Excitement	.14*	1.07 [0.99, 1.16]
Items (n = 8)
When I don’t understand things, I give up	.12^†	1.42 [0.93, 2.18]
When someone irritates me, I can’t control myself	.11^†	1.37 [0.89, 2.09]
It’s important not to show sadness in front of others	.12^†	1.35 [0.92, 1.98]
There is no better feeling than the rush I get when stealing	.15*	2.07 [1.18, 3.60]
I would rather steal or sell drugs than take a low paying job	.12^†	1.60 [0.97, 2.66]
It’s no big deal to go to jail, I’ll just do my time and get out	.11^†	1.47 [0.88, 2.47]
If I choose to commit a crime in the future, it is unlikely that I will get caught	.10^†	1.42 [0.90, 2.25]
It’s OK to get drunk or high, as long as you don’t do it around your kids	.28***	2.26 [1.47, 3.46]

†

p < .10. *p < .05. ***p < .001.

The scale was factorially derived and standardized with a sample of 1,320 adult probationers and inmates. Items are scored on a 4-point Likert-type scale. Higher scores reflect greater criminogenic thinking and can be interpreted with percentile ranks developed from the standardization sample. The CTP has been found to correlate positively with measures of psychopathy and aggressive/impulsive personality traits and negatively with measures of healthy personality functioning (Mitchell & Tafrate, 2012), and predict attrition from community justice programs (Mitchell et al., 2013). Alpha coefficients for the CTP total score were .94 in both the female and male samples in the current study.

Recidivism Data

Recidivism data were obtained from DRC case files and a check of records from the State of Connecticut Judicial Branch, and defined as a new arrest within 6 months of intake at the DRC. For purposes of data analysis, recidivism was dichotomously coded for each participant (0 = No arrest; 1 = Arrest). Overall, 22.2% (n = 48) of participants were rearrested (16.67% [n = 8] of female participants and 23.8% [n = 40] of male participants), with no significant gender differences in recidivism rates.

Results

CTP Scales by Gender

We first analyzed gender differences on the CTP with regard to prevalence. No gender differences emerged on the total score (Women: M = 109.90, SD = 20.36; Men: M = 109.10, SD = 21.25; t = −0.23, p = .82, d = .04) and few were noted on individual subscale scores. Women showed a greater tendency than men to demonstrate Parasitic/Exploitive thinking (M = 14.51, SD = 4.37 vs. M = 13.80, SD = 3.93, t = −2.86, p = .005, d = .47), whereas men showed a greater tendency to endorse the traditional criminogenic thinking pattern of Justifying (M = 10.39, SD = 3.41 vs. M = 9.41, SD = 3.38, t = 1.72, p = .06, d = .30).

In analyzing the relationship of the CTP to 6-month rearrest data by gender, two subscales emerged as gender-neutral (i.e., equally predictive of recidivism across gender): Poor Judgment and Justifying (see Table 1). One subscale, Disregard for Others, emerged as female-specific (i.e., predictive of recidivism for women but not men; see Table 2), whereas Demand for Excitement was male-specific (i.e., predictive of recidivism for men but not women; see Table 3). Neither of the remaining subscales nor the CTP total score predicted recidivism for either gender (see Table 4).

Table 4:

Logistic Regression Models: Criminogenic Thinking Scales

Model	n items (Cronbach’s α)	WomenAUC [95% CI]	MenAUC [95% CI]
Original CTP	W: 62 (α = .94); M: 62 (α = .94)	.64 [0.45, 0.83]	.58 [0.47, 0.69]
Gender-neutral	W: 7 (α = .77); M: 6 (α = .84)	.76 [0.53, 0.98]*	.65 [0.55, 0.76]**
Gender-specific	W: 12 (α = .72); M: 8 (α = .75)	.86 [0.74, 0.97]***	.67 [0.57, 0.77]***
Gender-responsive	W: 19 (α = .85); M: 16 (α = .90)	.86 [0.73, 0.99]***	.67 [0.56, 0.77]**

Note. AUC = area under the curve; CI = confidence interval; CTP = Criminogenic Thinking Profile.

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

Item-Level Analysis of CTP by Gender

Given the poor predictive validity of several of the CTP subscales, combined with the gender-specificity of others, we conducted an item-level analysis as an initial step to build optimal models for women and men, respectively. To establish baseline univariate statistical significance, we mapped the relationship of each CTP item to 6-month rearrest data by gender using the Spearman’s rho correlation coefficient. Given the small sample available for this pilot study, we opted to use a liberal inclusion threshold of p < .10 so as to safeguard against the premature exclusion of items. Table 1 presents items that empirically emerged as gender-neutral, with only 11.3% (n = 7) of the 62 CTP items meeting this definition. In contrast, 19.4% of items (n = 12) were female-specific (see Table 2), and 3.2% (n = 2) were female-salient (i.e., predictive of recidivism for both genders but significantly more so for women).⁴ Finally, as presented in Table 3, 12.9% of CTP items (n = 8) were male-specific. Notably, 52% (n = 33) of CTP items did not predict rearrest for women or for men, and were thus excluded from further analysis.

Gender-Informed Criminogenic Thinking Scale Development

As per above, items that did not demonstrate a significant association with recidivism were eliminated (p > .10; n = 33), as were items that evidenced extreme multicollinearity (rho > .70). In contrast to the original 62-item CTP, the final gender-responsive scales featured 19 items for women and 16 items for men. The gender-responsive model for women, for example, includes all items that are predictive of recidivism for the female subsample—including the gender-neutral items, female-specific items, and female-salient items (minus items that were multicollinear).⁵

As illustrated in Table 4, item scores from each collection (gender-neutral vs. gender-specific vs. gender-responsive) were simply added after which each aggregate score was entered into a separate logistic regression model. All developed scales showed acceptable internal consistency, with Cronbach’s alpha levels ranging from good to excellent (α = .72 to α = .90). Receiver Operating Characteristic (ROC) analysis was applied to examine the predictive validity of each model in forecasting new arrests over the 6-month follow-up period. The empirically derived gender-responsive scales predicted rearrest with a high degree of accuracy for women (area under the curve [AUC] = .86, p < .001) and a moderate degree of accuracy for men (AUC = .67, p = .002), outperforming the gender-neutral model (AUC = .76, p = .025, for women; AUC = .65, p = .005, for men) and representing a vast improvement over the original 62-item CTP that did not predict recidivism for either gender.⁶

Discussion

The present study used the scales and items of the CTP to explore gender differences in the relationship between criminogenic thinking and recidivism in a sample of justice-involved clients. Overall, the results of this pilot study suggest that (a) there are both gender-neutral and gender-specific thinking patterns and cognitions associated with recidivism, (b) reformulating the CTP items using a criterion-referenced approach and a gendered lens yields superior predictive ability over the original scale, and (c) when conceptualized in a gender-responsive manner, criminogenic thinking has a potentially greater relationship to the short-term recidivism of justice-involved women compared with men.

Gender-Neutral Patterns and Cognitions

Two of the CTP subscales and just over 11% of items are gender-neutral. At the subscale level, Justifying and Poor Judgment emerged as relevant predictors for both genders. These patterns can be conceptualized as types of Proactive criminogenic thinking within Walters’s (2019) overarching cognitive framework, and as Cognitive Immaturity according to Mandracchia et al.’s (2007) thematic structure. At the item level, gender-neutral predictors largely reflect a lack of interest in planning for the future, a tendency to see oneself as “smarter” than others in positions of authority, and justifications for substance use and antisocial behavior. It is interesting to note that engaging in criminal activity to support family members, often viewed as a contextual factor driving the lawbreaking activities of women (e.g., Chesney-Lind & Shelden, 2014), actually emerged as a strong justification for breaking the law across gender. Cognitions that reinforce excuses for criminal conduct and provide the illusion of averting negative consequences may be “universals” for putting justice-involved clients at risk for swift rearrest.

Gender-Specific Patterns and Cognitions

One CTP subscale, Disregard for Others, was predictive of rearrest for justice-involved women exclusively. This pattern can be conceptualized as Reactive criminogenic thinking within Walters’s (2019) overarching cognitive framework, and as a form of Egocentrism in Mandracchia et al.’s (2007) framework, indicating that emotion-driven thoughts triggered by situational or relationship challenges may put women most at risk for legal problems. At the item level, female-specific predictors of recidivism reflect a diverse array of cognitions clustered around four hypothesized themes: (a) Justification of substance use to cope with problems and boredom—coping with trauma and negative life events by abusing drugs is frequently noted in the gender-responsive literature on women’s pathways to crime (e.g., Daly, 1992; Scott et al., 2019); (b) dependence on people or institutions for financial support—echoing the theme of financial vulnerability (economic marginalization) among justice-involved women depicted in the feminist literature (e.g., Chesney-Lind & Shelden, 2014); (c) underestimating the impact of protective factors (e.g., education, participation in treatment programs) and risk factors (e.g., criminal history) on one’s future; and (d) revenge toward people perceived to have transgressed against them.

The revenge motive is consistent with relational-cultural theory and feminist pathways perspectives (e.g., Jones, 2011), positing that women place particular importance on relational bonds; by extension, a betrayal of such bonds is expected to be especially impactful. Hence, CBT interventions focused on positive coping strategies, consequential thinking, and relationship dynamics may be most promising for preventing rearrest among justice-involved women.

Interestingly, two gender-specific items for women were inversely associated with recidivism: For both “I am destined for greatness” and “I have more positive qualities than most people,” increases in scores were associated with a decreased likelihood of reoffending. In other words, perceiving positive qualities in oneself and envisioning a bright and meaningful future were protective for women, thus lowering their risk for rearrest. We hypothesize that these items represent a positive self-concept theme that is consistent with the gender-responsive literature on self-efficacy and self-esteem as a desistance factor in justice-involved women (Jones, 2011; Van Voorhis et al., 2010). By extension, cognitive interventions that boost self-confidence, self-compassion, and optimism may be particularly helpful for justice-involved women—especially among those with trauma-laden backgrounds. In contrast, such interventions may be contraindicated for justice-involved men. The above items were not significantly related to rearrest in the present sample of men, and other researchers have found that higher levels of self-esteem (i.e., grandiosity) are correlated with recidivism in male (or largely male) samples (e.g., Andrews et al., 1990; Wormith, 1984). Further research is recommended to elucidate the relationship between self-esteem and self-efficacy as moderators of criminal conduct for women and men, respectively.

One subscale, Demand for Excitement, was predictive of rearrest for men exclusively. This pattern can be conceptualized as a type of Reactive criminogenic thinking within Walters’s (2019) overarching cognitive framework, and as a form of Cognitive Immaturity in Mandracchia et al.’s (2007) framework, suggesting that impulsiveness related to excitement seeking is an important driver of criminality for men. The array of male-specific items predicting rearrest reflects four hypothesized themes: (a) emotional impulsivity and a tendency to give up or become easily frustrated when faced with challenging situations or people, (b) an unwillingness to reveal emotional vulnerability, (c) seeking excitement over conventional pursuits, and (d) indifference or a “no big deal” attitude regarding the justice system.

In contrast to the themes emerging with women, these echo a traditional masculine brand of antisociality depicted in the correctional literature characterized by externalizing features of impulsivity and stimulation-seeking (Jones et al., 2014; Skeem et al., 2011). As such, interventions designed to increase frustration tolerance, reduce impulsivity, and build awareness of how the consequences of justice involvement undermine life values may be particularly useful for combating recidivism among men.

Overall Predictive Utility of Gender-Informed Scales and Original CTP

All gender-informed models derived on the basis of the CTP items (whether gender-neutral, gender-specific, or gender-responsive) outperformed the full CTP. In fact, the CTP in its current 62-item form failed to predict rearrest for male or female subsamples at above chance levels—a finding clearly attributable to the fact that over 50% of the items were not associated with recidivism and consequently created “noise” in said models. This has implications for other criminogenic thinking instruments developed from a latent construct approach and suggests that many items may not be inherently related to criminal behavior. Assuming that the purpose (or least one of the purposes) of a measure is to predict some predefined criterion outcome, it is important to establish the predictive validity of items included on said measure.

That fact that so little gender-neutrality is evident among CTP items, combined with the high level of gender-specificity, also supports a gender-informed approach to developing criminogenic thinking assessment instruments. Indeed, the derivation of gender-responsive models resulted in two distinct parsimonious and reliable scales (with some overlapping gender-neutral items) that predicted rearrest to a high degree for women (AUC = .86) and a moderate degree for men (AUC = .67).⁷

Policy Implications

Although broad criminogenic content areas on risk/need assessments are relevant to both women and men, gender-specific nuances within those content areas should be integrated into assessment protocols if practitioners want to optimize prediction and, by extension, rehabilitation. The levels of predictive validity attained with the self-report gender-responsive models of criminogenic thinking in the current study (particularly the Female-Responsive model) are remarkable, particularly given that most risk assessment instruments that encompass multiple domains do not achieve such high accuracy. Adopting a gender-informed approach is likely to result in more effective assessment of justice-involved clients than a one-size-fits-all model.

In terms of guiding intervention, as D. J. Simourd and Olver (2002) lamented almost two decades ago, the criminogenic thinking assessment literature is virtually ignored in forensic practice, which is unfortunate because such instruments are readily available, free to use, and easily administered and scored. Adopting a gender-informed, criterion-validity approach in the development of self-report criminogenic thinking instruments has the potential to produce brief and efficient screening tools that may have greater appeal to real-world practitioners. Gender-responsive versions of the CTP (and other criminogenic thinking instruments) could potentially provide a useful adjunct to existing forensic assessment protocols in various justice settings.

Strengths, Limitations, and Future Directions

To the authors’ knowledge, the present study marks the first in the literature to examine gender differences in criminogenic thinking by applying an item-level methodology to optimize the prediction of recidivism. Additional strengths include the use of a reliable and empirically validated measure of criminogenic thinking in a sample of justice-involved women and men of diverse risk levels and ethnic backgrounds.

Nonetheless, the present results must be viewed as preliminary due to the limitations of the small sample size of women and the short recidivism follow-up period. The next step in this program of research is to obtain a larger sample of justice-involved women (and men) with a longer follow-up period to increase the generalizability of results. Ideally, we would also wish to explore criminogenic thinking in nonbinary and transgendered clients. Beyond the opportunity for replication, a larger sample would also allow for an examination of the extent to which crime severity and risk level moderate the relationship between criminogenic thinking and rearrest.

This study was also limited to the consideration of subscales and items from only one criminogenic thinking instrument. The addition of other measures (e.g., Psychological Inventory of Criminal Thinking Styles [PICTS], Criminal Sentiments Scale–Modified [CSS-M], and Measures of Criminal Attitudes and Associates [MCAA]) that contain a different constellation of criminogenic thinking patterns would likely uncover additional gender-specific items, providing not only improved risk indicators but also clarifying gender-specific cognitive targets for forensic CBT. Furthermore, increasing the number of items would allow for a blending of scale construction methodologies. Once a pool of predictive items has been identified, a factor analytic approach (latent construct) could then be applied to better elucidate different thinking patterns that could serve as potential treatment targets for female and male clients, respectively. Insufficient sample size, particularly of women, precluded a factor analysis of the gender-responsive scales derived in the present study.

A related vein of future research is to incorporate a more deliberate gender-informed methodology into the earliest stages of instrument development, which was not considered in the creation of the CTP or other existing criminogenic thinking instruments. A gender-informed approach could be adopted from the initial stages of item development by (a) incorporating hypothesized female-specific needs identified from the literature (e.g., substance-related coping, trauma history, relationship and parenting issues, financial dependence, low self-efficacy, low self-esteem) and (b) consulting with frontline staff to develop sample verbalizations that are potentially reflective of how criminogenic thinking manifests itself differently in justice-involved women. This ground-up approach could result in more nuanced items, which, by extension, could translate into more effective gender-responsive criminogenic thinking scales.

Conclusion

Writing about the landscape of correctional psychology, Steve Wormith noted that in the coming years, “Although we probably will still be asking about what works, answers to the more specific questions about for whom does it work, when does it work, and how does it work should become clearer” (Wormith et al., 2007, p. 888). The rich empirical literature he helped establish around RNR allows a subsequent generation of scholars to examine the nuances of for whom, when, and how the risk, need, and responsivity principles operate. The present study was conducted in the spirit of uncovering one of those nuances—an attempt to better understand the nature of gender differences in criminogenic thinking among justice-involved clients as it relates to risk. The results from this pilot project suggest there are both gender-neutral and gender-specific elements to the construct, and that capitalizing on these elements leads to better predictive utility. We expect the ensuing decade will see the further refinement of the nature of risk, need, and responsivity principles in assessment, case management, and treatment planning.

Footnotes

Authors’ Note:

We have no conflict of interest to disclose. For a copy of the Criminogenic Thinking Profile, contact Damon Mitchell at mitchelldam@ccsu.edu.

ORCID iD

Natalie J. Jones

Notes

Natalie J. Jones, PhD, is a senior evaluator at the Social Sciences and Humanities Research Council (SSHRC) of Canada and adjunct research professor in the Department of Psychology at Carleton University. Her primary research interests lie in the development of strength-based and gender-informed risk assessment and intervention strategies for justice-involved populations.

Damon Mitchell, PhD, is a clinical psychologist and professor of Criminology and Criminal Justice at Central Connecticut State University. As a criminal justice consultant, he has developed and delivered training workshops related to forensic assessment and treatment, and has conducted evaluations of criminal justice programs. He has published numerous journal articles and book chapters as well as a coedited book.

Raymond Chip Tafrate, PhD, is a clinical psychologist and professor in the Department of Criminology and Criminal Justice at Central Connecticut State University. His work is focused on problems of criminal behavior and anger in community settings. His most recent book is CBT With Justice-Involved Clients: Interventions for Antisocial and Self-Destructive Behaviors.

References

Andrews

D. A.

Bonta

(1995). The Level of Service Inventory–Revised. Multi-Health Systems.

Andrews

D. A.

Bonta

Hoge

R. D.

(1990). Classification for effective rehabilitation: Rediscovering psychology. Criminal Justice and Behavior, 17(1), 19–52. https://doi.org/10.1177/0093854890017001004

Andrews

D. A.

Bonta

Wormith

J. S.

(2004). Level of Service/Case Management Inventory: An offender assessment system. Multi-Health Systems.

Andrews

D. A.

Bonta

Wormith

J. S.

(2006). The recent past and near future of risk and/or need assessment. Crime and Delinquency, 52(1), 7–27. https://doi.org/10.1177/0011128705281756

Andrews

D. A.

Wormith

J. S.

(1984). Criminal sentiments and criminal behaviour [Programs branch user report]. Solicitor General Canada.

Andrews

D. A.

Wormith

J. S.

Kiessling

(1985). Self-reported criminal propensity and criminal behavior: Threats to the validity of assessment and personality [Programs branch user report]. Solicitor General Canada.

Barlow

D. H.

Farchione

T. J.

Sauer-Zavala

Latin

H. M.

Ellard

K. K.

Bullis

J. R.

Bentley

K. H.

Boettcher

H. T.

Cassiello-Robbins

(2017). Unified protocol for transdiagnostic treatment of emotional disorders: Workbook (2nd ed.). Oxford University Press.

Beck

A. T.

(1963). Thinking and depression: I. Idiosyncratic content and cognitive distortions. Archives of General Psychiatry, 9, 324–444. https://doi.org/10.1001/archpsyc.1963.01720160014002

Beck

A. T.

(1967). Depression: Causes and treatment. University of Pennsylvania Press.

10.

Beck

J. S.

(2011). Cognitive behavior therapy: Basics and beyond (2nd ed.). Guilford Press.

11.

Belknap

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

12.

Bloom

Owen

Covington

S. S.

(2003). Gender-responsive strategies for women offenders: Research, practice, and guiding principles for women offenders (NIC Accession Number 018017). https://nicic.gov/gender-responsive-strategies-research-practice-and-guiding-principles-women-offenders

13.

Bonta

Andrews

D. A.

(2017). The psychology of criminal conduct (6th ed.). Routledge.

14.

Chesney-Lind

Shelden

R. G.

(2014). Girls, delinquency, and juvenile justice (4th ed.). John Wiley.

15.

Daly

(1992). Women’s pathways to felony court: Feminist theories of lawbreaking and problems of representation. Southern California Review of Law and Women’s Studies, 2(11), 11–52.

16.

DiGiuseppe

Doyle

K. A.

Dryden

Backx

(2014). A practitioner’s guide to rational emotive behavior therapy (3rd ed.). Oxford University Press.

17.

Dowden

Andrews

(1999). What works for female offenders: A meta-analytic review. Crime & Delinquency, 45(4), 438–452. https://doi.org/10.1177/0011128799045004002

18.

Ellis

(1957). Rational psychotherapy and individual psychology. Journal of Individual Psychology, 13, 38–44.

19.

Ellis

(1962). Reason and emotion in psychotherapy. Lyle Stuart.

20.

Folk

J. B.

Stuewig

J. B.

Blasko

B. L.

Caudy

Martinez

A. G.

Maass

Taxman

F. S.

Tangney

J. P.

(2018). Do demographic factors moderate how well criminal thinking predicts recidivism? International Journal of Offender Therapy and Comparative Criminology, 62(7), 2045–2062. https://doi.org/10.1177/0306624X17694405

21.

Gendreau

Goggin

Law

M. A.

(1997). Predicting prison misconducts. Criminal Justice and Behavior, 24(4), 414–431. https://doi.org/10.1177/0093854897024004002

22.

Gendreau

Little

Goggin

(1996). A meta-analysis of adult recidivism: What works! Criminology, 34(4), 401–433. https://doi.org/10.1111/j.1745-9125.1996.tb01220.x

23.

Gobeil

Blanchette

Stewart

L. A.

(2016). A meta-analytic review of correctional interventions for women offenders: Gender-neutral versus gender-informed approaches. Criminal Justice and Behavior, 43(3), 201–322. https://doi.org/10.1177/0093854815621100

24.

Hannah-Moffat

(2006). Pandora’s box: Risk/need and gender-responsive corrections. Criminology and Public Policy, 5, 183–192. https://doi.org/10.1111/j.1745-9133.2006.00113.x

25.

Heimer

(1996). Gender, interaction, and delinquency: Testing a theory of differential social control. Social Psychology Quarterly, 59(1), 39–61. https://doi.org/10.2307/2787118

26.

Henning

Martinsson

Holdford

(2009). Gender differences in risk factors for intimate partner violence recidivism. Journal of Aggression, Maltreatment, & Trauma, 18(6), 623–645. https://doi.org/10.1080/10926770903103248

27.

Hubbard

D. J.

Pratt

T. C.

(2002). A meta-analysis of the predictors of delinquency among girls. Journal of Offender Rehabilitation, 34(3), 1–13. https://doi.org/10.1300/J076v34n03_01

28.

Jones

N. J.

(2011). Merging theoretical frameworks to inform risk assessment for the young female offender [Doctoral dissertation]. Carleton University Research Virtual Environment. https://curve.carleton.ca/system/files/etd/6acb828d-7c13-4493-b2ef-0d03ca67301c/etd_pdf/82710f5be28b212bfb24695fa1bc893e/jones-mergingtheoreticalframeworkstoinformriskassessment.pdf

29.

Jones

N. J.

Brown

S. L.

Wanamaker

K. A.

Greiner

L. E.

(2014). A quantitative exploration of gendered pathways to crime in a sample of male and female juvenile offenders. Feminist Criminology, 9(2), 113–136. https://doi.org/10.1177/1557085113501850

30.

Knight

Garner

B. R.

Simpson

D. D.

Morey

J. T.

Flynn

P. M.

(2006). An assessment for criminal thinking. Crime and Delinquency, 52(1), 159–177. https://doi.org/10.1177/0011128705281749

31.

Kroner

D. G.

Morgan

R. D.

(2014). An overview of strategies for the assessment and treatment of criminal thinking. In Tafrate

R. C.

Mitchell

(Eds.), Forensic CBT: A handbook for clinical practice (pp. 87–103). Wiley Blackwell.

32.

Latessa

Smith

Lemke

Makarios

Lowenkamp

(2009). Creation and validation of the Ohio Risk Assessment System: Final report.

33.

Mandracchia

J. T.

Morgan

R. D.

Garos

Garland

J. T.

(2007). Inmate thinking patterns: An empirical investigation. Criminal Justice and Behavior, 34(8), 1029–1043. https://doi.org/10.1177/0093854807301788

34.

Maruna

Lebel

T. P.

Mitchell

Naples

(2004). Pygmalion in the reintegration process: Desistance from crime through the looking glass. Psychology, Crime, and Law, 10(3), 271–182. https://doi.org/10.1080/10683160410001662762

35.

Mills

J. F.

Kroner

D. G.

Forth

A. E.

(2002). Measures of Criminal Attitudes and Associates (MCAA): Development, factor structure, reliability, and validity. Assessment, 9(3), 240–253. https://doi.org/10.1177/1073191102009003003

36.

Mills

J. F.

Kroner

D. G.

Hemmati

(2004). The Measures of Criminal Attitudes and Associates (MCAA): The prediction of general and violent recidivism. Criminal Justice and Behavior, 31(6), 717–733. https://doi.org/10.1177/0093854804268755

37.

Mitchell

Tafrate

R. C.

(2012). Conceptualization and measurement of criminal thinking: Initial validation of the Criminogenic Thinking Profile. International Journal of Offender Therapy and Comparative Criminology, 56(7), 1080–1102. https://doi.org/10.1177/0306624X11416197

38.

Mitchell

Tafrate

R. C.

Freeman

(2015). Antisocial personality disorder. In Beck

A. T.

Davis

D. D.

Freeman

(Eds.), Cognitive therapy of personality disorders (3rd ed., pp. 346–365). Guilford Press.

39.

Mitchell

Tafrate

R. C.

Hogan

Olver

M. E.

(2013). An exploration of the association between criminal thinking and community program attrition. Journal of Criminal Justice, 41(2), 81–89. https://doi.org/10.1016/j.jcrimjus.2012.09.003

40.

Mitchell

Wormith

J. S.

Tafrate

R. C.

(2016). Implications of risk-need-responsivity principles for forensic CBT. The Behavior Therapist, 39(5), 147–153.

41.

Morgan

R. D.

Kroner

D. G.

Mills

J. F.

(2018). A treatment manual for justice involved persons with mental illness: Changing lives and changing outcomes. Routledge/Taylor & Francis.

42.

O’Hagan

H. R.

Brown

S. L.

Jones

N. J.

Skilling

T. A.

(2019). The reliability and validity of the Measure of Criminal Attitudes and Associates and the Pride in Delinquency Scale in a mixed sex sample of justice-involved youth. Criminal Justice and Behavior, 46(5), 751–769. https://doi.org/10.1177/0093854818810459

43.

Olver

M. E.

Stockdale

K. C.

Wormith

J. S.

(2011). A meta-analysis of predictors of offender treatment attrition and its relationship to recidivism. Journal of Consulting and Clinical Psychology, 79(1), 6–21. https://doi.org/10.1037/a0022200

44.

Olver

M. E.

Stockdale

K. C.

Wormith

J. S.

(2014). Thirty years of research on the Level of Service Scales: A meta-analytic examination of accuracy and sources of variability. Psychological Assessment, 26(1), 156–176. https://doi.org/10.1037/a0035080

45.

Rice

M. E.

Harris

G. T.

(2005). Comparing effect sizes in follow-up studies: ROC Area, Cohen’s d, and r. Law and Human Behavior, 29(5), 615–620. https://doi.org/10.1007/s10979-005-6832-7

46.

Salisbury

E. J.

Van Voorhis

(2009). Gendered pathways: A quantitative investigation of women probationers’ paths to incarceration. Criminal Justice and Behavior, 36(6), 541–566. https://doi.org/10.1177/0093854809334076

47.

Scott

Brown

S. L.

Wanamaker

K. A.

(2019). Female offenders: Trends, effective practices, and ongoing debates. In Polaschek

Day

Hollin

(Eds.), The Wiley international handbook of correctional psychology (pp. 297–314). Wiley Blackwell.

48.

Shields

I. W.

Whitehall

G. C.

(1991, December). The Pride in Delinquency Scale [Paper presentation]. The Eastern Ontario Correctional Psychologists’ Winter conference.

49.

Simourd

D. J.

(1997). The Criminal Sentiments Scale–Modified and Pride in Delinquency Scale: Psychometric properties and construct validity of two measures of criminal attitudes. Criminal Justice and Behavior, 24(1), 52–70. https://doi.org/10.1177/0093854897024001004

50.

Simourd

D. J.

Olver

M. E.

(2002). The future of criminal attitudes research and practices. Criminal Justice and Behavior, 29(4), 427–446. https://doi.org/10.1177/0093854802029004005

51.

Simourd

Andrews

D. A.

(1994). Correlates of delinquency: A look at gender differences. Forum on Correctional Research, 6, 26–31.

52.

Skeem

J. L.

Polaschek

D. L. L.

Patrick

C. J.

Lilienfeld

S. O.

(2011). Psychopathic personality: Bridging the gap between scientific evidence and public policy. Psychological Science in the Public Interest, 12(3), 95–162. https://doi.org/10.1177/1529100611426706

53.

Steffensmeier

Allan

(1996). Gender and crime: Toward a gendered theory of female offending. Annual Review of Sociology, 22, 459–487. https://doi.org/10.1146/annurev.soc.22.1.459

54.

Sykes

G. M.

Matza

(1957). Techniques of neutralization: A theory of delinquency. American Sociological Review, 22, 664–673. https://doi.org/10.2307/2089195

55.

Tafrate

R. C.

Mitchell

Simourd

D. J.

(2018). CBT with justice-involved clients: Interventions for antisocial and self-destructive behaviors. Guilford Press.

56.

Tangney

J. P.

Stuewig

Furukawa

Kopelovich

Meyer

P. J.

Crosby

(2012). Reliability, validity, and predictive utility of the 25-item Criminogenic Cognitions Scale (CCS). Criminal Justice and Behavior, 39(10), 1340–1360. https://doi.org/10.1177/0093854812451092

57.

Taxman

F. S.

Rhodes

A. G.

Dumenci

(2011). Construct and predictive validity of criminal thinking scales. Criminal Justice and Behavior, 38(2), 174–187. https://doi.org/10.1177/0093854810389550

58.

Van Dieten

King

(2014). Advancing the use of CBT with justice-involved women. In Tafrate

R. C.

Mitchell

(Eds.), Forensic CBT: A handbook for clinical practice (pp. 327–353). Wiley Blackwell.

59.

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(3), 261–288. https://doi.org/10.1177/0093854809357442

60.

Walters

G. D.

(1995). The Psychological Inventory of Criminal Thinking Styles: I. Reliability and preliminary validity. Criminal Justice and Behavior, 22(3), 307–325. https://doi.org/10.1177/0093854895022003008

61.

Walters

G. D.

(2004). Predictors of early termination in a prison-based program of psychoeducation. The Prison Journal, 84(2), 171–183. https://doi.org/10.1177/0032885504265076

62.

Walters

G. D.

(2019). Criminal thinking: Theory and practice. In Polaschek

D. L.

Day

Hollin

C. R.

(Eds.), The Wiley international handbook of correctional psychology (pp. 637–653). Wiley Blackwell.

63.

Walters

G. D.

Lowenkamp

C. T.

(2016). Predicting recidivism with the Psychological Inventory of Criminal Thinking Styles (PICTS) in community-supervised male and female federal offenders. Psychological Assessment, 28(6), 652–659. https://doi.org/10.1037/pas0000210

64.

Walters

G. D.

Mandell

(2007). Incremental validity of the Psychological Inventory of Criminal Thinking Styles and Psychopathy Checklist: Screening version in predicting disciplinary outcome. Law and Human Behavior, 31(2), 141–157. https://doi.org/10.1007/s10979-006-9051-y

65.

Walters

G. D.

Schlauch

(2008). The Psychological Inventory of Criminal Thinking Styles and Level of Service Inventory–Revised: Screening version as predictors of official and self-reported disciplinary infractions. Law and Human Behavior, 32(5), 454–462. https://doi.org/10.1007/s10979-007-9117-5

66.

Wormith

J. S.

(1984). Attitude and behavior change of correctional clientele: A three year follow-up. Criminology, 22(4), 595–618. https://doi.org/10.1111/j.1745-9125.1984.tb00317.x

67.

Wormith

J. S.

Althouse

Simpson

Reitzel

L. R.

Fagan

T. J.

Morgan

R. D.

(2007). The rehabilitation and reintegration of offenders: The current landscape and some future directions for correctional psychology. Criminal Justice and Behavior, 34(7), 879–892. https://doi.org/10.1177/0093854807301552