Criminal Thinking, Age, Psychological Well-
Being, and Recidivism among Recently Released
Female Violent Offenders
Chapter 1: Introduction to the Study
In 2017, the United States incarcerated about 219,000, a number that had
increased nearly 50% since 2000, and almost 700% since 1980 (Seibold & FNI isenberg,
2018). Moreover, the incarceration rates for women have continuously risen in
comparison to men, with women as the fastest expanding population incarcerated in the
United States (Carson & Anderson, 2016). Indeed, the United States had the second
largest female prison rate in the world (64.6 per 100,000 people in the United States),
second only to Thailand (66.4; Krabbe & van Kempen, 2017).
While both men and women who have been incarcerated tend to come from low
educational and socioeconomic status, as well as have higher rates of childhood and adult
victimization levels and increased rates of mental illness and substance abuse, women
had a comparatively shorter criminal history and lower level of criminality than men
(Franke et al., 2019). Yet, prisons have tended to treat women and men in the same way,
but the focus on both the prison system and theories of criminality are rooted in
masculinity. In fact, female prisoners who were in prison often had different needs –
physical, emotional, and mental – which often have necessitated different treatment than
that given to male prisoners (Krabbe & van Kempen, 2017). The fact that prison systems
were male-oriented, in combination with the different needs of female prisoners, raises
questions as to the efficiency of current systems. In addition, while prison policy worked
along certain lines of logic – including serving a given sentence, rehabilitation,
resocialization, and perhaps even deterrence, the way to achieve these specific goals were
often different when dealing with men or women.
When women are released from prison, there are likewise significant
discrepancies between them and their male counterparts. Researchers indicated that the
recidivism rate for both men and women reentering the community was significant, with
almost two-thirds of released individuals rearrested within 3 years, and three-fourths
rearrested within 5 years (BJS, 2014). However, the reasons for recidivism were different
between the sexes. Van Ginneken (2015) found that incarcerated women had more
selfreported mental health and emotional problems related to prison, which had been
found to not only affect offenders’ psychological well-being in and out of prison
(including the risk of depression, substance abuse, and suicide), but had also been linked
to a reduction in recidivism post release (van Ginneken, 2015; Vrabel et al., 2019).
Recidivism has also been linked to criminal thinking, which involves two central
processes within an offenders’ thoughts: reactive and proactive criminal thinking. The
former involves reactions, indicating weak control over impulses and emotions, while the
latter suggested the ability to plan and be deliberate (Walters, 2107). Both of these
thought processes are important in gauging the likelihood of an individual to commit
antisocial behavior in the future. Researchers indicated that criminal thinking was
different for women and men, adhering to the gendered pathways model, which suggested
that social and traumatic experiences such as relationships, mental health, physical and
sexual abuse, and substance misuse were more significant in the development of female
offenders than males (Turanovic et al., 2015). My research was aimed at analyzing the
relationship between criminal thinking, age, psychological wellbeing, and recidivism
among recently released female violent offenders. The remainder of this chapter provides
an introduction to the dissertation as a whole. The next section offers a background on the
major ideas of this topic, as well as the related literature. Next, a problem statement is
offered, followed by the purpose statement. This study’s research questions, as well as the
concomitant hypotheses, are presented next, along with the theoretical framework for the
study, nature and significance of the study, and the definitions, assumptions, delimitations
and scope, and limitations of the study. Finally, this chapter ends with a summary.
Background
Criminal thinking and recidivism have been linked to one another in the extant
literature. Gavel and Mandracchia (2016) defined criminogenic thinking as the patterns of
cognitive events that were associated with criminal behavior. They explained that these
cognitive events allowed the development and maintenance of patterned criminal
behavior and addressed a gap in the literature related to the understanding of the specific
processes involved in criminogenic thinking and the development of criminogenic
thinking specifically. Criminal thinking was identified as important because it was a
major risk factor for criminal behavior among offenders (Vaske et al., 2016). In addition,
in a study by Walters and Lowenkamp (2016), the researchers used the Psychological
Inventory of Criminal Thinking Styles (PICTS) to predict recidivism in both male and
female released federal offenders. The researchers determined that the PICTS General
Criminal Thinking score was able to predict recidivism in both males and females in
terms of providing diagnostic information beyond what was provided through the use of a
comprehensive risk assessment procedure (Walters & Lowenkamp, 2016).
Vaske et al. (2016), on the other hand, explored the issue of gendered differences
in criminal thinking. The researchers explained that there was a lack of consensus in
research on whether there were gendered differences in criminal thinking and whether
females had the same level of criminal thinking as males. The results of the Vaske et al.
(2016) study were that 26% of the items analyzed were different between genders, but
that male and female probationers were just as likely to exhibit antisocial attitudes. Based
on these findings, Vaske et al. (2016) concluded that it was important that researchers did
not assume that criminal thinking assessments were gender-neutral or that similar results
would be yielded between genders.
Further, there have been crucial differences found between former inmates in their
pathways to, experiences in, and release from prison on the basis of factors that included
age and mental health. Within prison, van Ginneken (2015) explained that there was
evidence that imprisonment caused a criminogenic rather than a rehabilitative effect and
that an individual’s prison experience may have affected their outcomes post release.
Based on the results of interviews, the researcher concluded that even among prisoners in
the same prison environment, there were variety of patterns and differences in
psychological adjustments. Gemeda (2017) explained that serious crimes and recidivism
may have been directly related to psychopathic personality traits. Furthermore,
psychopathic personality traits may have also helped to explain the effect of antisocial
behaviors and mental health and its relation to recidivism.
Age had also been shown to significantly impact how criminal thinking and other
risk factors and experiences affected the likelihood of recidivism. Walters (2020) found
that age significantly moderated the relationship between criminal thinking and perceived
expectations of legal punishment and age had been used as a predictive variable in some
models aimed at predicting recidivism within the U.S. criminal justice system. However,
other research has revealed no significant variation in the association between recidivism
and criminal thinking when demographics were not treated as control variables. Thus,
there is a lack of clarity in existing literature concerning how age impacted recidivism
and recidivism risk factors (Benson & Harbison, 2020; Folk et al., 2018; Vaske et al.,
2016; Walters, 2020).
Researchers have found that any paradigms for understanding recidivism, or
models for reintegration into society, were premised on male understandings and ignored
the differences that men and women developed, criminally or otherwise, differently
(Gobeil et al., 2016). Crewe et al. (2017) noted that there was a lack of literature related
to the practices and effects of imprisonment on women, and that there were distinct
differences that existed between male and female prisoners, particularly differences
related to the problems experienced by each group. The issues that were identified as
most important among women, when compared to male inmates, were loss of contact
with family members, power, autonomy and control, and psychological well-being and
mental health and issues related to trust, privacy, and intimacy.
Prior research had linked recidivism and criminal thinking among various
populations of previously incarcerated individuals. However, there remains a lack of
understanding pertaining to associations between age, gender, criminal thinking, and
recidivism, and the intersections of these factors among currently or previously
incarcerated women (Walters, 2018; Walters, 2019). This study addressed the gap in the
literature, thereby contributing more knowledge to field of female-based offenders and
hopefully aid in creating female-centered models of reintegration.
Problem Statement
The general topic that was explored in this study was the criminal thinking and
recidivism of women incarcerated in the United States. Researchers had suggested that
women in prison reported emotional and mental health problems that may have been
related to the prison environment (Caulfield, 2016; van Ginneken, 2015). The mental
health and psychological wellbeing of incarcerated individuals is important because
mental illness, including the reduction of feelings of despair and isolation were important
in reducing the risk of suicide, depression, and substance abuse among prisoners as well
as reduced recidivism post-release (Molleman & van Ginneken, 2014; Pimlott Kubiak et
al., 2015; Scott et al., 2015; Scraton, 2016; van Ginneken, 2015; Vrabel et al., 2019).
The specific topic that was explored in this study was the relationship between criminal
thinking, psychological wellbeing, age, and recidivism among previouslyincarcerated
female violent offenders that had been released from prison in the last 2 years using
Yochelson and Samenow’s criminal thinking theory as a predictor of recidivism (Folk et
al., 2016; Pimlott Kubiak et al., 2015; Pantalone et al., 2018). There is a lack of literature
on the use of criminal thinking theory and recidivism among violent female offenders;
based on this gap in literature, additional research was needed to understand the
relationship between recidivism, age, and criminal thinking among previously-
incarcerated female offenders, particularly as gender, age, and criminal thinking were all
factors related to and potential predictors of recidivism (Collica-Cox & Furst, 2018).
While some literature existed on the use of criminal thinking theory as related to
recidivism, there was a lack of literature with a focus on the use of Yochelsen and
Samenow’s criminal thinking theory and assessment on female offenders. Similarly,
although there were studies on gendered differences between prisoners, the gap existed in
the application of criminal thinking theory for female offenders after their release into the
community (Adams et al., 2017; Crewe et al., 2017; Walters & Lowenkamp, 2016).
Purpose of the Study
The purpose of this quantitative research study was to analyze the relationship
between criminal thinking, age, psychological well-being, and recidivism among recently
released female violent offenders through the use of Yochelsen and Samenow’s criminal
thinking theory. Addressing this gap was important to understand the potential gender
differences and applicability of criminal thinking theory for females and how it related to
the recidivism of female offenders that had reentered the community. Additionally,
addressing this gap was important because in research on the criminal thinking of
offenders, female offenders were often overlooked, as was the impact of age (Link &
Oser, 2018; Pantalone et al., 2018; van Ginneken, 2015; Vaske et al., 2016). This study
also contributed to understanding the role of criminal thinking on the recidivism of
formerly incarcerated female violent offenders.
Research Questions and Hypotheses
The following research questions and hypotheses were developed to guide the
study:
Research Question 1 (RQ1): Is there a statistically significant difference between
the criminal thinking scores of recently released female violent offenders and the
normative control sample of non-offenders as measured by the Psychological Inventory
of Criminal Thinking Styles (PICTS) assessment?
H01: There is no statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control
sample of non-offenders as measured by the Psychological Inventory of Criminal
Thinking Styles (PICTS) assessment.
Ha1: There is a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control
sample of non-offenders as measured by the Psychological Inventory of Criminal
Thinking Styles (PICTS) assessment.
Research Question 2 (RQ2): Is there a relationship between the age of first
incarceration and the score measured by the Psychological Inventory of Criminal
Thinking Styles (PICTS) assessment.
H02: There is no relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
Ha2: There is a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
Research Question 3 (RQ3): Is there a statistically significant association between
the age of female offenders and the number of incarcerations?
H03: There is no statistically significant association between PICTS scores related
to age and number of incarcerations.
Ha3: There is a statistically significant association between PICTS scores related
to age and number of incarcerations.
Research Question 4 (RQ4): Is there a relationship between the mental well-being
as measured by the Measure of Psychological Well-Being (MPWB) and criminal thinking
as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment?
H04: There is no relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
Ha4: There is a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
Theoretical Framework
The theoretical framework that was used in this study was based on Yochelson
and Samenow’s (1976; 1977) theory on criminal thinking. Criminal thinking theory, also
referred to by researchers as criminogenic thinking, was the concept that criminals
thought differently and had different personalities than noncriminal (Yochelson &
Samenow, 1976; 1977). The three main categories of criminogenic thinking are
criminogenic thinking patterns, automatic errors of thinking, and problems in the thinking
process between idea and execution (Mandracchia et al., 2015). Researchers also have
suggested that there is a link between general thinking error and specific thinking patterns
that perpetuates problematic behaviors (Mandracchia et al., 2015). Understanding
criminal thinking is important because there was an association between recidivism and
psychopathic characteristics which was associated with criminogenic thinking
(Mandracchia et al., 2015).
Criminal thinking was useful and beneficial to this study to understand the
differences in results between offenders and nonoffenders. Criminal thinking could be
used to understand the nature of criminal cognition based on the context of incarceration
and how these may have been related to criminal behavior, particularly as related to
recidivism. Specifically, whether the environment of incarceration reinforced antisocial
behavior and criminal acts and how routines may have played a role in such behavior
(Morgan et al., 2015). There was limited literature available regarding the application of
Yochelson and Samenow’s theory of criminal thinking (1976; 1977) to incarcerated and
recently released female violent offenders specifically. Therefore, this study also
contributed to the existing research by adding to the limited literature available.
Nature of the Study
I used a correlational cross-sectional research design, which was an appropriate
choice since the objective of the study was to measure variables and analyze them using
statistical analysis to explain phenomena (Mustafa, 2011). More specifically, this study
used a set of research questions and associated hypotheses to be tested, consistent with
quantitative psychology (Else-Quest & Shibley Hyde, 2016). In order to determine
association of the results to the characteristics of female violent offenders, a control group
was used that consisted of female nonoffenders.
This study explored the relationship between criminal thinking, psychological
wellbeing, age, and recidivism among previously-incarcerated female violent offenders
who had been released from prison in the last 2 years using Yochelson and Samenow’s
criminal thinking theory. Because this data was not currently available, the sources of
data for this study included the criminal thinking assessment, the well-being assessment,
and online flyer survey completed by participants. The Psychological Inventory of
Criminal Thinking (PICTS) questions regarding reoffense and incarceration were used to
measure past recidivism, and other questions were used to assess criminal thinking.
PICTS was selected as a measure of recidivism because there was evidence that the
PICTS General Criminal Thinking score was able to predict recidivism in both males and
females (Walters & Lowenkamp, 2016). The Measure of Psychological Well-Being
(MPWB) was the well-being assessment that was used (Choi et al., 2014). The MPWB
was selected due its validity in relation to comparable instruments and the direct, succinct
nature of the seven items included in the scale.
The primary target population for this study included the adult female violent
exoffenders in Central Texas. The ideal sample size for this population was 128
participants. The minimum sample size for this study was 82, which was computed using
power analysis. The required number of samples was determined through power analysis.
Power analysis was conducted through G*Power software. The assessment was
administered online once the participant accepted the terms of the study and gave
informed consent agreement in the beginning of the survey. The online administration of
the assessment allowed for standardization and removed potential bias associated with
inperson administration. The platform that was used to administer the assessments was
Survey Monkey. Survey Monkey was selected due to feasibility and the low-cost
associated with the platform. Participants involved in this study were asked to complete a
consent form and were informed of their ability to discontinue or refuse to answer any
part of the assessments at any time. Due to the sensitivity of this data, the data collected
in this study was kept in a password protected file on my computer. In responding to
Research Question 3, participants were also asked whether they had reoffended at any
time and how many times they had been incarcerated as measures for recidivism. Linear
regression was used in this study to analyze the connections between measures from the
assessment on criminal thinking and measures related to the questions on recidivism, to
identify and statistically significant associations. Linear regression analysis was also used
to identify statistically significant findings from the assessments as well as associations
between the results of each assessment.
Definitions
Criminal thinking: A collection of intervening variables that can link crucial
independent variables from core criminological theories to various categories of criminal
conduct (Yochelson & Samenow, 1976; 1977).
Deterrence: Methods that help to discourage offenders from committing further
criminal activity (Calvi & Coleman, 2000).
Ex-offenders: Persons who have been released from prison and have returned to
the community (James, 2015).
Gendered pathways to crime model: This model posits that females are impacted
by different factors than males in their development of offending. These factors include
substance misuse, mental health issues, sexual and abuse, substance misuse, and
relationship challenges (Chesney-Lind & Pasko, 2004).
Inmates: Any individuals who are confined in an institution for rehabilitation,
typically prison (James, 2015).
Intersectionality: A theory that originated in black feminist thought,
intersectionality considers multiple aspects of identity, difference, and inequality,
including gender, race, class, etc., in the theoretical or analytical approach of the study
(Else-Quest & Shibley Hyde, 2016).
Recidivism: The repetition of or the return to criminal behavior by the same
person, which results in the offender’s re-arrest, re-conviction, or return to prison
(Chenane et al., 2014).
Reentry: A generalized term that is used to indicate issues, programs, and services
that are connected to the transition that offenders experience when they move from prison
back to their community (BJS, 2016).
Assumptions
There were several assumptions that undergirded this proposed study. The first
assumption was that all the respondents would respond to the questionnaires openly and
honestly. Moreover, there was an assumption that all participants would be able to
understand the language and questions used in the online flyer survey. An additional
assumption was that participants would have a personal understanding of factors that
contributed to criminal thinking and/or recidivism. Finally, this study would proceed
under the assumption that the survey instruments proposed in this study would measure
validly and reliably.
Scope and Delimitations
There were several factors that delimited this study and its scope. First, the focus
of the proposed study was on female ex-offenders in Texas, which excluded male
exoffenders, as well as offenders outside of this geographic location. Thus, the results of
this study were generalizable to the population of adult female violent ex-offenders in the
geographic region considered in the study. Second, the sampling of this proposed study
relied on prospective participants who were willing and available to participate in the
study. Third, this study was delimited to analyzing the concepts of criminal thinking,
psychological well-being, and recidivism.
Limitations
A limitation of the study would be the generalizability of the results of the study
to other populations as the sample would be restricted to an accessible group of recently
incarcerated female violent offenders. However, a purposeful sample was calculated to
maintain the likelihood of the generalizability of the data. A minimum sample size had
also been identified in consideration of maintaining the feasibility of the study. An
additional limitation of this study was that there had not been a criminal thinking
assessment designed specifically for women and there was a lack of consensus as to
whether gendered differences existed in terms of criminal thinking, with evidence that
criminal thinking assessments differed between males and females (Vaske et al., 2016).
For this reason, the results of this study was specific to female offenders and can not
necessarily be generalized to male offenders.
Significance
Applying Yochelson and Samenow’s criminal thinking theory (1976; 1977), this
study explored the relationship between criminal thinking, psychological wellbeing, age,
and recidivism among previously-incarcerated female violent offenders that had been
released from prison in the last 2 years using Yochelson and Samenow’s criminal
thinking theory. By analyzing the criminal thinking of and psychological wellbeing of
recently released female offenders, this study added to the limited literature on female
offenders and also contributed to the literature by investigating the connections between
criminal thinking, psychological wellbeing, age, and gender as potential predictors of
recidivism. Addressing this gap was important because in research on the criminal
thinking of offenders, female offenders were often overlooked, as was the impact of age
(Link & Oser, 2018; Pantalone et al., 2018; van Ginneken, 2015; Vaske et al., 2016). As
reflected in the background section, there were similarly a lack of understanding and
evidence related to the factors associated with criminal thinking among female offenders
and how these factors differed from those of male offenders (Crewe et al., 2017; Folk et
al., 2016; Link & Oser, 2018; Pantalone et al., 2018; van Ginneken, 2015; Vaske et al.,
2016; Walters & Lowenkamp, 2016).
In addition, the results from this study may help to inform policy related to the
role of criminal thinking and psychological wellbeing on the risk of recidivism among
previously-incarcerated women, as well as informed intervention programs aimed to
reduce criminal thinking and recidivism among violent female offenders specifically.
This was important given that the recidivism rate for both men and women reentering the
community is significant, with almost two-thirds of released individuals rearrested within
3 years, and three-fourths rearrested within 5 years (BJS, 2014). Moreover, such results
were crucial given the current paradigms used for understanding recidivism and the
models for reintegration into society, which were premised on male understandings and
ignored the differences that men and women developed, criminally or otherwise,
differently (Gobeil et al., 2016). The results of this study helps bring to light the need for
gender-specific interventions and approaches.
Summary
This quantitative study aimed to explore the relationship between criminal
thinking, psychological wellbeing, age, and recidivism among previously-incarcerated
female violent offenders who had been released from prison in the last 2 years (Folk et
al., 2016; Pimlott Kubiak et al., 2015; Pantalone et al., 2018). Based on the gap in
literature, additional research was needed to understand the relationship between
psychological wellbeing and criminal thinking, age, and the recidivism of female violent
offenders, particularly as age, gender, and criminal thinking were all factors related to and
potential predictors of recidivism (Collica-Cox & Furst, 2018). The theoretical
framework used in this study was based on Yochelson and Samenow’s theory on criminal
thinking, and the group of interest in this study were female violent offenders in Texas.
The proposed criminal thinking assessment used in this study was the PICTS, while the
proposed well-being assessment was the MPWB. Chapter 2 will review the extant
literature on the topics addressed by and in this study.
Chapter 2: Literature Review
Introduction
This research centered on criminal thinking and recidivism among women who
were previously incarcerated in the United States. Given the persistently high rates of
recidivism in U.S. prisons (BJS, 2014), I aimed to address several gaps in extant
literature, which included the connection between psychological wellbeing, criminal
thinking, age, gender, and recidivism; the use of criminal thinking theory among violent
female offenders; and the application of criminal thinking theory for female offenders
after their release into the community. Considering these research gaps, the purpose of
the quantitative research study analyzed the relationship between criminal thinking,
psychological wellbeing, age, gender, and recidivism among recently released female
violent offenders with Yochelson and Samenow’s criminal thinking theory (1976; 1977).
In Chapter 2, I first describe my literature search strategy. Then, the theoretical
framework of the study, the theory of criminal thinking, is explored. Subsequently, the
U.S. prison system is then discussed. I then discuss reentry into society following time in
prison and recidivism, including subsections on predictors of recidivism and means of
preventing recidivism, followed by a discussion of psychological well-being among
prisoners. A summary concludes the chapter.
Literature Search Strategy
To find articles relevant to the present study, EbscoHost and Google Scholar
databases were searched. The following words and phrases were searched to locate
articles: criminal thinking, criminal thinking theory, recidivism, female inmates, women,
incarceration, deterrence, ex-offenders, and reentry. Most articles included in this review
were written within the past 5 years to ensure their current relevance to the research
focus. However, some older seminal and theoretical works were included as well to
inform the historic and theoretical basis of the study.
Theoretical Framework
The theoretical framework that guided this study was Yochelson and Samenow’s
theory on criminal thinking, or criminogenic thinking (1976; 1977). Criminogenic
theorists contend that criminals and noncriminals differ cognitively. Criminal thinking
theorists generally categorized the ways that criminally divergent thinking differs in three
ways: criminogenic thinking patterns, automatic errors of thinking, and problems in the
thinking process between idea and execution (Mandracchia et al., 2015). Criminal
thinking theory was a useful and effective theory for lending insight into the causes of
criminal behavior, recidivism, gender-related differences among incarcerated individuals,
and other research aimed at explaining/examining criminality (Yochelson & Samenow,
1976; 1977).
There were a number of studies that implemented criminal thinking theory to
guide their research. Criminal thinking theory informed research into mental health issues
within offenders (Folk et al, 2016; Mandracchia et al., 2015;). Its connection to emotional
intelligence had also been investigated (Westfall, 2019). Criminal thinking had been
associated with stressful life events and culture (Link & Oser, 2017). Mandracchia et al.
(2015) and Ziegler-Hill et al. (2017) investigated the role psychopathy had in criminal
thinking. Folk et al. (2018) revealed no significant variation between recidivism and
criminal thinking on the basis of demographic variables, including gender. Rudin et al.
(2019) contested the COMPAS model and more variables needed to be considered.
Though research has shown there are differences in criminal thinking based on gender
(Benson & Harbison, 2020), most research into criminal thinking and gender has focused
on comparing men and women (Vaske et al., 2016).
In my study, criminal thinking theory was used as a guide to frame female
experiences and perspectives of criminal thinking, psychological wellbeing, age, and
recidivism. Constructing an accurate picture of the patterns of behavior associated with
criminal thinking was necessary to analyze ways to help offenders successfully reenter
society (Ziegler-Hill et al, 2017; Mandracchia et al., 2015). Cognition that resulted in
multiple incarcerations would need to be different from nonoffenders, so this study
utilized criminal thinking theory to inform its investigation into the criminal thinking of
female violent offenders.
Criminal thinking theory was the best choice for this investigation into the
relationships between the PICTS score, age, and incarceration. Criminal thinking posits
that cognitive differences would be present, which I sought to assess through the
investigation into the first research question. This investigation also investigating the
relationship between criminal thinking, age, first incarceration as well as total
incarcerations to build an image of what kind of criminal thinking patterns were present
in female violent offenders. Criminal thinking theory helped explain repeated negative
behavior by exploring cognitive differences between offenders and noncriminals that lead
to incarceration.
Review of the Literature
The U.S. Prison System
The United States prison system is a complex and multi-faceted entity comprised
of both private and public institutions. Approximately 1,435,500 people were
incarcerated in state and federal prisons at the end of 2019 (Kang-Brown et al, 2020). Of
those incarcerated, approximately 231,000 were women (Kajstura, 2019). Thus,
approximately 16% of U.S. prisoners were female in 2019.
While this figure was very high in relation to the prison populations of other
countries with similar economic and political contexts, it has decreased significantly since
a peak in 2007. In 2007, the prison population was approximately 7,339,600 (Kaeble et
al., 2016). Since then, it had decreased annually by approximately 1%. The total prison
population at the beginning of 2015 was the lowest since 2003 due to public policy
changes, reform efforts, and several other factors (Figure 1). However, when considering
that one in 36 individuals living in the U.S. were incarcerated at this lowpoint, questions
still aroused about the efficacy of the system and the number of prisoners that were repeat
offenders (Kaeble et al., 2016).
Figure 1
Estimated Total Population Under the Supervision of the U.S Adult Correctional System,
by Correctional Status, 2000-2014.
Note. From Kaeble et al., 2016.
The U.S. prison populace was by no means distributed evenly where geography
was concerned. Rather, approximately 50% of the prison populace once resided in seven
jurisdictions (Kaeble et al., 2015). More specifically, by the end of 2014, 50% of the
prison populace resided in Texas, California, Georgia, Florida, Pennsylvania, Ohio, and
Michigan (Kaeble et al., 2015). Thus, a significant focus in recent research had been
examining why certain regions had significantly higher rates of crime and incarceration
than others.
There were many points of contention associated with the U.S. prison system
related to aspects such as the potential for behavioral correction/rehabilitation, safety,
sentencing, and mental health care (Kaeble et al., 2015; Sawyer & Wagner, 2019). One
significant point of contention pertained to the system of requiring bail payments that
could vary significantly depending on which judge determined the bail amount. In many
cases, a rich or well-connected person could be released immediately because they had
the resources to make bail, while a poor person may have been held for weeks, months, or
even years waiting for their sentencing or to go to trial. Individuals being held in U.S.
prisons for not being able to pay their bail amount made up a significant majority (76%)
of the U.S. prison population (Sawyer & Wagner, 2019). With the average bail amount
being set at $10,0000, many lower and even middle-class citizens were detained for a
significant amount of time based solely on their lack of financial resources and assets.
Another significant point of contention was the number of nonviolent offenders
who were incarcerated for years at a time. Many critics of the current system noted that it
treated those with mental illnesses and substance abuse issues punitively rather than
seeking to heal or rehabilitate them (Green & Jackson, 2017; Sawyer & Wagner, 2019).
As of 2019, one in five individuals that were incarcerated in the U.S. were in prison for
drug offenses and violations; most of these individuals were housed in state prisons
(Sawyer & Wagner, 2019). Another large subsect of non-violent individuals incarcerated
in the U.S. were in prison due to probation and parole violations, rather than the crime
that led to their probation/parole. Further, sometimes non-violent offenders or partners of
violent offenders were threatened with prosecution and/or jail time for refusing to testify
or incriminate their spouses (Green & Jackson, 2017). Non-violent offenders were not
selected for this study as they were not likely to have the same deeply ingrained criminal
thinking as a repeat violent offender.
"Prison-industrial complex" was a term that was used by many to describe the
U.S. prison system (Green & Jackson, 2017). The term described the infrastructure and
network that propelled the prison system to be a multi-billion-dollar industry that
countless businesses profited from. In the 1970s, many public policies that were "tough
on crime" led to an increase in the disproportionate incarceration of many marginalized
citizens, primarily people of color, the poor, and the uneducated.
Many prisons promoted programs and efforts to rehabilitate prisoners, decreasing
criminality and, by extension, the number of prisoners that required housing in
correctional institutions. Prison programs designed to decrease recidivism in the U.S.
and/or improve the well-being of prisoners could broadly be characterized as (a)
rehabilitation, (b) education, or (b) vocational training (Dick, 2018). Effective prison
programs had been shown to decrease recidivism, which, in turn, decreased prison costs
(Dick, 2018). However, there was limited data pertaining to the effectiveness of programs
designed to reduce recidivism in the U.S., particularly where programs for specific prison
populations (i.e. women, those who served time for specific crimes, individuals with
substance abuse issues) were concerned (Dick, 2018). Further, the profits of many
correctional institutions were also dependent on the number of prisoners they housed;
thus, some private prisons and correctional institutions had a financial motive to keep
rehabilitation efforts from being exceedingly successful (Green & Jackson, 2017).
Criminal Thinking
Connections between certain mental illnesses and criminogenic thinking had been
found in recent research. Mandracchia et al. (2015) sought to analyze psychopathy and
criminogenic thinking among male incarcerated offenders in the Mississippi Department
of Correction. The researchers found that increased indicators of primary and secondary
psychopathy significantly predicted increased overall and specific subtypes of
criminogenic thinking. Based on these findings, Mandracchia et al. (2015) suggested that
prison-based treatment programs should address psychopathic personality characteristics
and criminogenic thinking. Thus, implications and insights from criminal thinking
research could contribute to improved incarceration and mental health treatment
outcomes.
Folk et al. (2016) similarly found evidence of a relationship between criminal
thinking and mental health. More specifically, the researchers assessed the effectiveness
of a cognitive-behavioral criminal thinking intervention intended to be self-administered
to inmates living in segregated housing. A total of 273 inmates participated in the Taking
a Chance on Change (TCC) intervention. Analysis of intervention data revealed a
significant decrease in criminal thinking among most participants; further, 48 participants
experienced a significant decrease in disciplinary infractions after completing the TCC
program. This finding coincided with the additional finding that decreased reactive
criminal thinking were a predictor of decreased disciplinary infractions. Findings from
Folk et al.’s (2016) study highlighted how interventions and programs aimed at
addressing criminal thinking could also effectively reduce criminal behavior, violent
outbursts, and other unfavorable conduct among current and former inmates.
Criminal thinking was analyzed by Morgan et al. (2015) using two studies, one
used college student participants and a second used inmate participants, to understand
whether levels of criminal thinking were fixed or fluid across situational contexts. The
hypothesis of the researchers were that criminal thinking and antisocial attitudes would be
related to increased proximity to a criminal act. The researchers found that the results
between the two studies were generally consistent between the two participant groups.
Based on the findings of the study, Morgan et al. (2015) concluded that additional
research was needed to understand criminal cognitions over time and whether criminal
cognitions change based on different environmental factors.
Not all recent criminal thinking research had been conclusive. For instance,
Morgan et al. (2015) studied criminal thinking among two separate populations to
examine similarities and differences. In one study, a sample of college students
participated; in another otherwise identical study, inmates participated. The researchers
hypothesized that criminal thinking and antisocial attitudes would be related to increased
proximity to a criminal act. The researchers found that the results between the two studies
were generally consistent between the two participant groups. Based on the findings of
the study, Morgan et al. (2015) concluded that additional research was needed to
understand criminal cognitions over time and whether criminal cognitions change based
on different environmental factors.
A recent study conducted by Vaske et al. (2016) offered potential explanations for
the lack of conclusive research findings: “One explanation for these inconsistent results
was that it may have been difficult to conceptualize and operationalize criminal thinking”
(p. 2). The researchers went on to delineate further sources of confusion surrounding the
notion of criminal thinking: “The conceptualization of criminal thinking was often broad
in the literature. There were also questions surrounding how the concept should be
operationalized and what was the best way to assess criminal thinking (i.e., interview vs.
self-report, close-ended questions vs. open-ended questions)” (p. 2). Further, the accuracy
of measures of criminal thinking varied across gender which could contribute to null
associations (Vaske et al., 2016).
There was a lack of consensus in extant literature regarding whether age-related
and gendered differences were apparent where criminal thinking was concerned (Benson
& Harbison, 2020; Folk et al., 2018’ Vaske et al., 2016; Walters, 2020). Age was
examined as a potential moderating variable of the association between general criminal
thinking and perceived expectations of legal punishment by Walters (2020). The
researcher found that age did significantly moderate the studied relationship; the
relationship between general criminal thinking and perceived expectations of legal
punishment was not significant at age 17 but was significant at age 21 and beyond
(Walters, 2020). Age had also been used as a predictive variable in models, such as the
Correctional Offender Management Profiling for Alternative Solutions (COMPAS)
model, that, like criminal thinking scores, was used to predict recidivism within the U.S.
criminal justice system and guide decisions pertaining to “judicial bail, parole,
sentencing, lending decisions, credit scoring, marketing, and access to social services” (p.
2).
However, in contrast to these findings, another recent study conducted by Folk et
al. (2018) revealed no significant variation in the association between recidivism and
criminal thinking on the basis of demographic variables, including age, race, gender,
psychological well-being, and education level. Further, Rudin et al. (2019) contested a
premise that guided the development of the COMPAS model; the researchers contended
that while there may have been a relationship between age and the likelihood of
recidivism as predicted by the COMPAS model, the model was not linear as the model
developers purported it to be and was significantly influenced by other demographic
variables such as race and gender (Rudin et al., 2019). These findings largely indicated
that while relationships may have existed between age, criminal thinking, recidivism,
and/or likelihood of reoffending, these relationships were not well-understood or salient
when demographic controls were not in place.
Considering the identified research gap, Vaske et al. (2016) sought to delineate
whether criminal thinking was conceptualized differently based on gender and whether
men and women possessed the same latent criminal thinking levels. A total of 375
probationers participated by filling out a survey. Upon analyzing the survey data, the
researchers found mixed results. There appeared to be gender differences in survey scores
which were significant, as 26% of item variance was tied to gender. However, some
evidence indicated that the survey questions were more valid for participants of a certain
gender due to DIF. Thus, while Vaske et al. (2016) found evidence of gender-related
differences in criminal thinking, the researchers also noted the need for further research to
determine whether notions of criminal thinking were inaccurately biased towards male
criminal thinking and experiences.
Gender and criminal thinking had also been studied based on specific crimes that
were committed (Benson & Harbison, 2020). Using the PICTS, Benson and Harbison
(2020) recently compared the scores of women and men who were convicted of and
incarcerated for white collar crimes. Analysis of participants’ responses revealed low
criminal thinking scores for participants of both genders when no controls were put in
place. However, when factors which predicted criminal thinking were controlled by the
researcher, women had significantly higher scores on all three of the scales that were
considered than men. Despite this finding, other factors associated with risk of criminality
and personal needs had stronger effects than gender. Benson and Harbison’s (2020)
research demonstrated how criminal thinking could differ based on gender, as well as the
importance of controlling for other factors that predicted criminal thinking when seeking
to understand how criminal thinking differed based on gender.
It should be noted that most of the recent research which centered on the
relationship between gender and criminal thinking was framed as a comparison of women
and men, or female and male offenders (Vaske et al., 2016). There was a significant body
of literature that pertained to criminal thinking solely among males/male prisoners, and a
significant lack of research with a central focus on criminal thinking among
females/female prisoners. During one such study, Westfall (2019) recently explored the
association between criminal thinking and emotional intelligence among female prisoners
in Iowa. The PICTS and Emotional Quotient Inventory were used. The results of
Westfall’s (2019) research highlighted how experiences, factors, and contexts that were
specific to female prisoners could impact criminal thinking. For instance, there were
disproportionate rates of sexual assault and childhood sexual abuse among female
prisoners; these experiences could detrimentally affect social-emotional development and
indirectly increase reactive criminal thinking. Further, the lack of available research on
criminal thinking styles among female prisoners paralleled a lack of recidivism
prevention programs in the U.S. prison system that were effectively tailored to help
female prisoners (Westfall, 2019).
Another recent study conducted by Link and Oser (2017) centered on how
stressful life events and culture influenced criminal thinking among 418 African
American women that participated in the B-WISE project. The findings indicated that in
certain contexts and among certain populations, criminal thinking could be a maladaptive
approach to cope with stressors. In particular, coping with gendered racism, loss of social
network, financial challenges, and other stressors that resulted from going to prison or
being on probation could lead to criminal thinking among African American women; in
turn, use of criminal thinking as a coping mechanism could then lead to recidivism (Link
& Oser, 2017).
Age
Age had also been shown to significantly impact how criminal thinking and other
risk factors and experiences affect the likelihood of recidivism. Walters (2020) found that
age significantly moderated the relationship between criminal thinking and perceived
expectations of legal punishment, and age had been used as a predictive variable in some
models aimed at predicting recidivism within the U.S. criminal justice system. However,
other research had revealed no significant variation in the association between recidivism
and criminal thinking when demographics were not treated as control variables. Thus,
there were a lack of clarity in the existing literature concerning how age impacted
recidivism and recidivism risk factors (Benson & Harbison, 2020; Folk et al., 2018;
Vaske et al., 2016; Walters, 2020). There was a lack of consensus in extant literature
regarding whether age-related and gendered differences were apparent where criminal
thinking was concerned (Benson & Harbison, 2020; Folk et al., 2018’ Vaske et al., 2016;
Walters, 2020). Age was examined as a potential moderating variable of the association
between general criminal thinking and perceived expectations of legal punishment by
Walters (2020). The researcher found that age did significantly moderate the studied
relationship; the relationship between general criminal thinking and perceived
expectations of legal punishment was not significant at age 17 but was significant at age
21 and beyond (Walters, 2020). Age had also been used as a predictive variable in
models, such as the model, that, like criminal thinking scores, were used to predict
recidivism within the U.S. criminal justice system and guided decisions pertaining to
“judicial bail, parole, sentencing, lending decisions, credit scoring, marketing, and access
to social services” (p. 2).
However, in contrast to these findings, another recent study conducted by Folk et
al. (2018) revealed no significant variation in the association between recidivism and
criminal thinking on the basis of demographic variables, including age, race, gender,
psychological well-being, and education level. Further, Rudin, Wang, and Coker (2019)
contested a premise that guided the development of the COMPAS model; the researchers
contended that while there may have been a relationship between age and the likelihood
of recidivism as predicted by the COMPAS model, the model was not linear as the model
developers purported it to be and was significantly influenced by other demographic
variables such as race and gender (Rudin et al., 2019). These findings largely indicated
that while relationships may have existed between age, criminal thinking, recidivism,
and/or likelihood of reoffending, these relationships were not well-understood or salient
when demographic controls were not in place.
Recidivism and Rehabilitation
Among inmates that were not sentenced to life imprisonment, recidivism and
rehabilitation were significant concerns. Recidivism described a criminal reoffending or
committing another crime, following their release from prison (Carr, Baker, & Cassidy,
2016). High recidivism rates served as an indicator that the current prison system was
somewhat ineffective, as the goals of the prison system were to deter criminal activity
and rehabilitate individuals so that they did not re-offend after their release (Carr et al.,
2016; Mitchell, Cochran, Mears, & Bales, 2017; Stemen, 2017). Recidivism had been a
common topic in recent literature and research among populations of violent offenders as
well as those incarcerated for non-violent offenses, such as drug violations and
prostitution. Certain non-violent offenses, such as prostitution, were a reason for
incarceration that disproportionately affected women (Pantalone et al., 2018). Despite
women being more likely than men to reoffend after being incarcerated for certain
nonviolent offenses, there remained a lack of available data and programs which centered
on recidivism among female inmates.
Researchers aimed to understand why individuals who had been previously
incarcerated re-offend, and which ex-inmates were most likely to offend, in the hopes of
developing solutions to reduce rates of recidivism (James, 2015). Many prisoners were
not serving life sentences and were eventually released; thus, recidivism was a concern
that applied to most prison populations (Alper et al., 2018; Carlson, 2018; Carr et al.,
2016).
As with most research on U.S. prisoners, recidivism research had largely centered
on male experiences. Thus, while gender differences were easily identifiable in terms of
recidivism rates, gender differences associated with the variables and factors that were
closely tied to recidivism were less clear (Olson et al., 2016). Recidivism was measured
in both reconviction and reimprisonment rates (Fazel & Wolf, 2015). While both types of
recidivism were considered in this study, reimprisonment was the focus, as
reimprisonment could have a more significant impact on well-being and social integration
than being convicted a second time without reimprisonment.
When considering recidivism and recidivism research, it was imperative to
consider the potential cumulative effects of multiple challenges and comorbidities which
affected the experiences of those who were previously imprisoned. Pantalone et al. (2018)
sought to examine the unmet social service and mental health needs of women who were
previously incarcerated. The study subjects lived in two different cities in Alabama and
had been diagnosed with HIV. The researchers conducted semi-structured interviews with
participants to gain insight into their experiences. Analysis of the interview data revealed
that there was a significant and influential lack of services related to planning for prison
release and life after prison. Further, a small number of organizations were offering post-
release services but were unduly burdened with helping a large number of ex-inmates.
Post-release adjustment and avoiding recidivism were significantly dependent on the
availability of substance abuse and/or mental health treatment, as well as social support
(Pantalone et al., 2018).
Substance abuse and mental health treatment options were available at many
prisons across the country, and may have included prescription treatment, group and/or
individual therapy, art programs, and other approaches designed to improve the lives of
inmates with substance abuse and/or mental health issues. However, a lack of resources
in some prisons limited the scope of treatment approaches and programs (Dick, 2018;
Pantalone et al., 2018). Social support could be understood as the various individuals and
community resources that helped and comforted individuals, particularly in times of
distress. Findings from Pantalone et al.’s (2018) research helped to delineate the key
elements of avoiding recidivism and effectively re-adjusting for women who were
previously incarcerated and were dealing with one or more health and wellness
comorbidities, namely substance abuse disorder and mental health disorders which
significantly impacted the daily life of inmates (i.e. PTSD, schizophrenia, bipolar
disorder).
Olson et al. (2016) compared recidivism and prison release risk factors among
male and female inmates. The researchers differentiated between rearrests for violent and
non-violent crimes in their calculated rates. Logistic regression of data gathered from the
Illinois State Police and Illinois Department of Corrections was used, as it remained
difficult to find complete recidivism data for a given region that could be gathered from a
single source in the U.S. Data from 3,014 female and 23,520 male individuals were
included in the analysis. Upon conducting a regression analysis with the collected data,
Olson et al. (2016) found some significant differences associated with gender. Types of
arrest, time served, and prior incarceration were found to be significantly related to
female and male recidivism, but these associations differed in strength. Men and women
were both more likely to re-offend if they had a history of violent arrests. Previous arrests
for selling drugs predicted female recidivism for violent offenses, but not male
recidivism. Conversely, males were less likely to re-offend violently based on a higher
number of previous arrests for drug possession, but this effect did not apply to females.
Access to substance abuse treatment was associated with decreased recidivism among
both female and male inmates. Findings from Olson et al.’s (2016) research highlighted
how risk factors for violent and non-violent recidivism differed, as well as similarities
and differences between factors that contributed to and reduced recidivism among men
and women.
When correctional institutions sought to prevent recidivism through targeted
interventions and programs, gender was not always a consideration. However, it remained
unclear whether recidivism interventions were more effective when they were tailored to
the gender and gendered experiences of current or former inmates. Gobeil et al. (2016)
recently examined and compared gender-neutral and gender-informed approaches to
recidivism prevention interventions. A meta-analysis was conducted across 37 studies, 38
effect sizes, and 22,000 currently or formerly incarcerated women. Data analysis revealed
that participating in some form of a recidivism prevention intervention was associated
with a 22-35% higher likelihood of a subject not reoffending. Further, it was revealed that
after excluding studies with small effect sizes or low-quality methods, gender-informed
interventions were significantly more effective than gender-neutral interventions. Thus,
Gobeil et al. (2016) suggested the importance of consideration of female experiences and
perspectives when developing interventions aimed at reducing recidivism among women.
Tools to predict recidivism had been tested among male and female inmates
(Walters, Glenn, & Lowenkamp, 2016). A recent study conducted by Walters et al.
(2016) involved testing the Psychological Inventory of Criminal Thinking Styles (PICTS)
as a means of predicting recidivism. A sample of 14,519 female and 81,881 male
offenders who had served time and were on probation responded to a survey including the
PICTS items, and the results were subsequently analyzed. Analysis of the data revealed
that some scores and items (reactive, proactive, and criminal thinking) predicted
recidivism at six, 12, and 24 months post-release. Controlling for prior arrests and age
revealed small and moderate effect sizes, respectively. Further, general criminal thinking
scores from the assessment predicted recidivism more effectively among both female and
male ex-inmates than overall scores on the PICTS (Walters et al., 2016). The results of
Walters et al.'s (2016) research did not reveal significant gender differences on PICTS
scores among those included in the sample; however, it was important to note that the
sample of female ex-inmates was significantly smaller than the male sample and thus,
may have been less reflective of the average experiences and likelihood of recidivism
among women who were previously incarcerated.
Mitchell et al. (2017) used regression analysis methods to examine how
effectively serving time in prison reduced recidivism among those convicted of drug
offenses. Data reflected the recidivism and sentencing data of individuals convicted of
felony drug charges in Florida. Regression analysis revealed no significant reduction in
recidivism as a result of time served in prison. Those sentenced for the most minor drug
offenses were not less likely to re-offend based on serving time in prison. Findings from
Mitchell et al’s (2017) research called into question the implications and effectiveness of
time in prison as a deterrent for non-violent felony reoffenders.
While there remained a lack of comprehensive data on female recidivism in the
U.S., some were available. A report compiled by Alper et al. (2018) indicated that based
on data from 2005, 35% of female prisoners were rearrested during the first year
following their release, compared to 45% of male prisoners. Nine years after the initial
data was collected, discrepancies based on gender narrowed with 24% of male prisoners
and 21% of female prisoners arrested (Alper et al., 2018). Thus, based on this limited
cohort data recidivism was lower among women, but that the gap in recidivism rates
decreased over time after ex-inmates were released. The researchers’ findings were based
on a nine-year cohort study which included 401,288 prisoners who were released from
prisons in 30 states.
A meta-analysis of adult recidivism in the U.S. was recently conducted by
Katsiyannis et al. (2017). The researchers noted a significant need for the study, as the
last comprehensive meta-analysis of U.S. recidivism before their research was conducted
in 1996. Their meta-analysis included all instances of re-offense reported after any
criminal was released from prison between the years of 1994 and 2015. Upon analyzing
publicly available data, the researchers found that age, antisocial personality scales,
distress, criminogenic needs, criminality within the family, gender, family rearing, risk
scales, antisocial behavioral history, substance abuse, and social achievement were
significant predictors of recidivism in the U.S. Findings from Katsiyannis et al.’s (2017)
research highlighted the multitude of factors, including gender, that impacted recidivism
rates; it was important to note that some of these factors, including family rearing,
disproportionately impacted female offenders.
Another recent study conducted by Western et al. (2015) highlighted significant
hardships and stressors experienced by recently released inmates. Namely, the researchers
emphasized the difficulties of the social re-integration process following time in prison.
Panel data from 122 individuals who were recently released from prison was sourced
from the Boston Reentry Study. The data was collected using a comprehensive survey of
post-prison reentry experiences. Upon analyzing the data for items relevant to the topic of
the study, Western et al. (2015) found significant unemployment issues among
respondents. Over 50% of respondents were unemployed, while over 66% received
public assistance. Some respondents were more prone to experience problems related to
housing stability, unemployment, and severance of family ties; specifically, those with
mental illnesses and substance abuse issues were the most prone to experience these
challenges. Further, many participants relied on mothers, sisters, grandmothers, and other
female relatives for financial assistance in the months following their release. Other
challenges revealed by the survey included persistent feelings of isolation, material
insecurity, and anxiety. The researchers concluded by emphasizing how the struggle of
readjusting socially to society combined with material insecurity could lead to significant
stress and isolation among recently released inmates.
Challenges and issues experienced by ex-inmates as they re-enter society,
including those highlighted by Western et al. (2015) contributed to recidivism. The sense
of isolation and other challenges of reintegration made it difficult for many ex-inmates to
re-connect to their old lives and potentially positive influences. Issues of unemployment
and lack of opportunity could make some crimes more appealing, particularly for those
who could not acquire legal employment because of a felony conviction. Further,
disconnection from social influences and isolation during and after time in prison also
contributed to decreased motivation to avoid the lifestyle and consequences of
criminality. For this reason, Western et al. (2015) contended that success after ex-inmates
reentered society should not be based solely on whether or not they re-offended
(recidivism), but rather, how they socially reintegrated into their community: "Our focus
on social integration broadens the definition of “success” after incarceration. In contrast
to the usual focus on recidivism, a successful transition from prison in our analysis
involves attaining a basic level of material and social well-being consistent with
community membership” (p. 1515).
Many factors were tied to recidivism or had been studied to determine how they
were linked to recidivism (Agan & Makowsky, 2018; Reagan, 2017; Zgoba & Salemo,
2017). Certain mental health conditions, such as antisocial personality disorder, and
substance abuse were closely linked to recidivism (Gemeda, 2017; Westerberg,
McCrady, Owens, & Guerin, 2016). Hardship, be it personal or financial, could have also
led to desperation and additional criminality following release from prison. Similarly, a
lack of access to resources or job opportunities because one’s criminal past could also
lead to recidivism. Experiences of recidivism were complex, multi-faceted, and were still
not well-understood when considering the existing body of research on the subject.
Predictors of recidivism
High rates of recidivism around the world have necessitated increased research on
factors that predicted recidivism (Ahmed, 2015). A significant portion of extant research
on predictors of recidivism centered on the U.S. prison system, given that the U.S. had
the highest rate of recidivism among industrialized global nations (Sellers, 2016).
However, research from other parts of the globe remained useful where insights about
psychology and personality-related factors were concerned, as these constructs could
have been analyzed independently from regional characteristics and variables. A
qualitative study conducted by Ahmed (2015) was aimed at determining which factors
predict recidivism based on four predictive constructs: stigma, personality, prison, and
discrimination. Participants were ex-prisoners that served time in Nigerian prisons.
Narrative analysis of interview data revealed that stigma, prison, and discrimination had a
significant influence on criminal recidivism. Conversely, personality and personality
factors were not found to significantly influence recidivism. The results of Ahmed’s
(2015) research suggested that experiences and circumstances may have affected
recidivism as much as, if not more than, personality traits and characteristics. Lauch,
Hart, and Bresler (2017) studied recidivism alongside the results of a treatment program
for offenders who previously committed intimate partner violence.
Domestic violence was an issue that disproportionately affected women in the U.S.;
approximately 1 in 3 women, and 1 in 10 men, were victims of domestic violence
(Huecker & Smock, 2019). Lauch et al. (2017) analyzed data that reported which
offenders out of a sample of males who had been convicted of intimate partner violence
against their female partners (n = 202) had completed the AMEND-Emerge-based
program and which had not. Program data was then compared to archival corrections data
to determine which program participants reoffended following their completion or
incompletion of the program. Within the sample, data revealed a relatively low rate of
recidivism (22.28%). Those who completed the program had lower rates of recidivism
than those who did not, though the difference was small. Comparison of demographic
factors to the recidivism and program data revealed that the program may have addressed
some participants' needs more effectively than others; namely, young African Americans
who were unmarried may have not benefited as significantly from the AMEND-
Emergebased program as individuals who represented other demographics (Launch et al.,
2017). Findings from Launch et al.'s (2017) study highlighted the importance of ensuring
efforts to prevent recidivism were effective among all relevant demographics and prison
populations, including both women and men who perpetrated intimate partner violence.
Robertson et al. (2018) studied gender-specific predictors of recidivism. The sample that
was analyzed represented 10,827 men and women who were previously convicted of a
DUI in Mississippi. Recidivism scales were used to predict recidivism within a twelve-
month interval after the DUI conviction. Upon using quantitative methods to analyze the
data, Robertson et al. (2018) found that different factors predicted recidivism depending
on the gender of subjects. Heavy consumption of alcohol and criminal history predicted
recidivism among men, while current or past substance abuse issues were a predictor of
recidivism among women. Predictors of recidivism that applied to both men and women
were found to be past DUI arrest, physical consequences of substance abuse, and driving
behaviors. Findings from Robertson et al.’s (2018) research revealed significant
differences in factors that predicted recidivism among men and women; however, it was
important to note that some of these factors, such as driving behavior, were specific to ex-
offenders that were arrested for one or more DUI(s).
In a related study conducted 2 years earlier, Robertson, Gardner, Walker, and
Tatch (2016) examined DUI recidivism based on adherence to an anti-DUI intervention.
Multiple risk factors were also considered. Data from the Mississippi Alcohol Safety
Education Program (MASEP) and state records were analyzed. Analysis of the data
revealed that individuals who demonstrated effective adherence to the MASEP program
presented a significantly lower likelihood of recidivism within the three years after
program completion. Among those who completed the program, recidivism was more
likely among younger individuals, African Americans, and those with low education
levels. Adherence to the program was more common among African American and older
program participants. Gender was not found to be a significant predictor within the
context of Robertson et al.’s (2016) research. Findings from Robertson et al.’s (2016)
research highlighted the significant connection between adherence to treatment programs
and recidivism among DUI offenders, as well as the importance of ensuring that
treatment programs were equally effective among all relevant demographics and
subpopulations.
Harris, Boccaccini, and Rice (2017) also studied recidivism among offenders
based on the crime they committed in the past. Their focus was on examining sexual
deviance and psychopathy as recidivism predictors among individuals who were
previously convicted of a sexual offense. Various scales and field measures were used to
assess psychopathy and sexual deviance among 687 individuals who were released after
being convicted of a sexual offense. Data analysis revealed that scores on a revised
psychopathy checklist and anti-social personality disorder diagnoses were predictors of
recidivism where a violent-sexual offense was committed, but not when an offense was
solely violent or sexual in nature (Harris et al., 2017). No evidence suggested that
individuals with high levels of both sexual deviance and psychopathy would re-offend at
a higher rate than the average. The results of Harris et al.'s (2017) research highlighted
the complex relationship between psychological deviance, psychopathology, and
recidivism among individuals convicted of sex crimes.
Research on predictors of recidivism among offenders who had perpetrated family
violence was recently conducted by Millsteed and Coghlan (2016). Because there was a
disproportionate prevalence of female domestic violence/abuse victims and male
domestic violence/abuse perpetrators (Huecker & Smock, 2019), the implications of the
study largely reflected recidivism committed by male perpetrators against female victims.
The authors noted the need for the study based on the lack of tools available which could
be used to predict recidivism among those who had previously committed domestic
assault or violence. Logistic regression analysis and modeling were used to test various
predictors. Consistent with past research findings, a history of prior incidents, a breach of
a family violence order, or having pending criminal charges for violating a family
violence order were found to be predictors of recidivism among domestic violence
perpetrators. Substance abuse was also found to be significantly related to recidivism,
though this predictor was not specific to perpetrators of domestic violence. Millsteed and
Coughlan (2016) concluded by suggesting the need for further piloting and evaluation of
tools to predict recidivism among sub-populations of ex-inmates in Victoria.
In another study with a similar focus, Farzan-Kashani and Murphy (2017) studied
whether long-term criminal recidivism was predicted by anger issues among men who
were previously convicted of partner violence. The period that was studied was the eight
years following offenders' release from prison after being charged with domestic
violence. The sample included 132 men who took part in a violence prevention treatment
program. Findings from analysis of correctional and program data revealed that the more
significant an offender’s anger problems were, the more likely they were to have a high
number of general violence changes and protection orders against them. Further,
recidivism was predicted by high Anger Expression and low Anger Control among those
in the sample (Farzan-Kashani & Murphy, 2017). Thus, while violence prevention
programs may have been effective to reduce violence and recidivism perpetrated by some
domestic violence offenders, certain subgroups, namely those who struggled to express
and control their anger appropriately, may not have experienced success in these
programs. Further research was necessary to determine whether issues of anger
expression and/or control were connected to violence and recidivism prevention among
female offenders.
Psychological Well-Being and Recidivism
The psychological health and well-being of inmates were a commonly discussed
topic in extant literature about the U.S. prison environment (Bar-on, 1988; Franke et al.,
2019; Keogh et al., 2017). Significant concerns had been raised for years due to the high
prevalence of mental health disorders among inmates and the limited availability of
psychological support and resources within prisons. Some cognitive and mental health
conditions, including posttraumatic stress disorder, attention deficit hyperactivity
disorder, antisocial personality disorder, and learning disabilities, were more common
than others within prison populations (Stanford & Muhammad, 2017). Beyond specific
conditions, issues and episodes of substance use, self-harm, and suicide were also more
prevalent among prison populations than the general populace (Franke et al., 2019).
Certain mental health conditions and disorders make both female and male
inmates more likely to be incarcerated more than once (Gemeda, 2017). Gemeda (2017)
recently examined whether psychopathic personality disorder, a condition characterized
by antisocial and/or violent behavior, mediated recidivism among former inmates. The
researcher involved a sample of 196 adults who were convicted and incarcerated for
multiple crimes. Semi-structured interview methods were used to gather data. The author
used structural equation modeling to also consider the implications of drug abuse, and the
influence of associates and social exclusion. Analysis of the data revealed that recidivism
was indirectly influenced by drug abuse, the influence of associates, and social exclusion
by means of psychopathic personality disorder (Gemeda, 2017). Thus, while drug abuse,
the influence of associates, and social exclusion were not found to directly affect
recidivism in this context, one or more of these issues coupled with a diagnosis of
psychopathic personality disorder were closely linked to recidivism. Findings from
Gemeda's (2017) research highlighted the complex interplay of factors that could
contribute to recidivism.
A review of recent developments and findings pertaining to mental healthcare in
prison environments by Franke et al. (2019) revealed significant challenges and research
gaps. Where challenges were concerned, several mental health conditions and negative
effects were found among prisoners. Substance use disorder treatment outcomes were
found to be promising in the prison context, but psychological therapies were not found
to be effective over an extended period. Franke et al. (2019) concluded by noting
significant research gaps and needs for future research where prison mental healthcare
was concerned; namely, research on female prisoners, those with multiple comorbidities,
and testing of innovative diagnostic and therapeutic treatment methods.
Though mental health resources could be scarce in prison, some mental health
interventions, workshops, and programs have been tested. Keogh et al. (2017) recently
examined the outcomes and implications of a one-day mental health workshop
implemented among male inmates at a prison in Ireland. Semistructured phone interviews
were conducted with ten prisoners who had participated in the workshop. Quantitative
evaluation was conducted in a different phase of the study. Analysis of the qualitative
data revealed that the participants had primarily positive feedback to share regarding how
the workshop affected their mental health and well-being. Notably, the opportunity for
prisoners who had time to reflect on their mental health and strategize towards improving
their well-being in a structured and supportive environment was a unique experience that
stayed with participants long after their participation. Keogh et al. (2017) also highlighted
common issues and challenges which affected the delivery of mental healthcare in the
prison environment, noting that identifying those with mental health issues was not useful
if appropriate services could not be provided. The researchers also noted that typically
interventions were medical in nature and focused on the distribution of medication.
Further, while the identification of mental health problems among prisoners who were on
remand had improved, there was little in the way of treating those who were incarcerated
or likely to have had mental health issues.
Results and insights from Keogh et al.'s (2017) study highlighted how research on
mental healthcare interventions and treatment in the prison setting should be compatible
and realistic within the scope of possibility for the care prisons could provide.
Determining an ideal or favorable treatment approach for prison populations was
pointless if the populations that demonstrated favorable results in a research setting would
not realistically benefit from such high-quality treatment on a large scale. Thus, research
on prison mental healthcare should have involved prioritization of contextual
considerations including but not limited to prison resources, public policy, and the
conditions/changes necessary to implement suggested changes and initiatives effectively.
Recent research had revealed associations between the psychological well-being
of ex-inmates and their likelihood of reoffending (Bales et al., 2017; Schaftenaar et al.,
2018). Mental illness could further complicate the process of re-entry into society;
exinmates with serious mental illnesses were less likely to find stable employment and
social networks following time in prison (Duwe & Johnson, 2016). Further, some
individuals who had difficulty finding mental healthcare and resources may have
experienced poor mental healthcare outcomes after they no longer had access to
healthcare in prison (Lamberti, 2016).
Bales et al. (2017) recently studied recidivism in relation to mental illness among
prisoners in U. S. jails. A cohort of 200,889 inmates who were released from prisons in
Florida between 2004 and 2011 was studied. Of those included in the sample, the
proportion of inmates diagnosed with a serious mental illness and/or mental health
diagnosis was compared to those without a diagnosis/mental illness. Analysis of
recidivism data for the cohort revealed that ex-inmates with a mental illness were
significantly more likely to re-offend in comparison to those without a mental illness.
Schaftenaar et al. (2018) studied recidivism among individuals who were
sentenced to spend time in a Forensic Psychiatric Hospital (FPH) as a result of being
charged with a crime in the Netherlands. The researchers examined the two years
following the release of 111 patients. Rates of criminality following discharge for those
who were sentenced to time at an FPH due to criminality were compared to rates of
criminality among those who were receiving care as usual, as well as a control group that
reflected average recidivism rates. Data analysis revealed that the recidivism rate among
those sentenced to time at an FPH was significantly lower than the average rate of
recidivism represented by the control group (15.6% and 46.5%, respectively). Further,
patients who received continued contact and support after being discharged from an FPH
were even less likely to re-offend. Findings from Schaftenaar et al.’s (2018) research
emphasized the importance of effective mental healthcare and continued support for
reducing recidivism among those convicted of crimes.
Similarly, Lowder et al. (2016) studied recidivism among those who were
previously sentenced to treatment in a Mental Health Court (MHC). The authors noted
the rapid expansion of MHC systems across the U.S. due to the significant number of
mentally ill individuals who would otherwise be processed through criminal courts. A
sample of data from MHC patients was compared to data from individuals receiving
treatment as usual at the same facility to assess recidivism. Individual characteristics,
demographics, and process factors were also considered alongside recidivism. Analysis of
the data led to the finding that while MHC participants spent fewer days in jail than the
control group, they did not have fewer convictions or charges. Even more significant
decreases in jail time following treatment were found among those who experienced
cooccurring substance use, MHC graduation, and longer MHC participation length.
Findings from Lowder et al.’s (2016) research emphasized how MCHs may have offered
a solution to reducing recidivism, as well as removing mentally ill offenders from prison
environments which may have exacerbated mental illness.
Lamberti (2016) contended that collaboration between criminal justice and mental
health organizations hold the solution to preventing a significant portion of criminal
recidivism. Specialty probation, mental health courts, conditional release programs, and
other approaches for addressing criminality perpetrated by individuals with mental
illnesses all entailed the use of legal leverage to motivate adherence to state-prescribed
treatment. There were some conflicting findings in research regarding the efficacy of
such approaches, though a growing body of literature seemed to indicate that the close
involvement of criminal justice authorities in mental illness treatment and justice
approaches increased rates of recidivism (Lamberti, 2016). To ensure that the structure
and administrators of mental health treatment approaches to criminal justice were
effective, Lamberti (2016) concluded that best practices from both the fields of criminal
justice and mental healthcare should be heeded and that a stepwise process consisted of
"engagement, assessment, planning and treatment, monitoring, problem-solving, and
transition” (p. 1210) should be followed when mental healthcare professionals collaborate
with criminal justice and law enforcement organizations.
Preventing Recidivism
Due to the far-reaching concerns and implications related to recidivism, many
prison leaders and researchers have aimed to develop solutions to reduce and prevent
recidivism (Carr et al., 2016; Collica-Cox & Furst, 2018). Various approaches and
programs have been developed to prevent recidivism. Some focused on specific prison
populations (i.e. women, drug users, mentally ill prisoners), while others centered on
offering education and other resources. There remained a significant lack of recidivism
prevention programs that were tailored to help female inmates (Collica-Cox & Furst,
2018).
A key component of preventing recidivism was ensuring individuals who were
released from prison had the highest possible likelihood of successfully rejoining society.
Thus, many programs aimed at preventing recidivism were conceptualized as “reentry”
programs. James (2015) described common programs implemented in U.S. prisons that
were aimed at improving reentry into society. Typically, there were three phases
associated with an offender reentering society. There were programs in place for while an
offender was incarcerated aimed at preparing offenders for their release and connecting
them to services. Finally, there were programs designed to assist offenders long-term at
integrating into their communities by providing support and supervision focused on job
training and placement, drug and mental health treatment, and house assistance.
Regardless of how effectively programs designed to curtail recidivism were,
obstacles and challenges were unavoidable. Collica-Cox and Furst (2018) studied one
such program aimed at preventing recidivism among female inmates. The researchers
noted the significant need for recidivism prevention programs aimed at female prison
populations, as their needs were often overlooked because there were significantly more
male prisoners incarcerated across the country. The program, Parenting, Prison & Pups
(PPP), described by the authors falls into the category of programs described by James
(2015) that were implemented before inmates were released back into society. The PPP
program was developed to reduce recidivism by incorporating parenting curriculum and
animal-assisted therapy.
Despite the PPP program being evidence-based and mimicking elements of
successfully implemented programs, Collica-Cox and Furst (2018) highlighted certain
challenges associated with program implementation that would be inevitable. Namely,
issues with program implementation were usually related to “gaining sponsorship,
successfully navigating large bureaucracies, obtaining all of the necessary levels of
approval from multiple agencies, and negotiating facility schedules” (p. 112). However,
effective time management, patience, policy adherence, dedication, and flexibility could
help to mitigate such issues. In terms of implementing the PPP program specifically,
Collica-Cox and Furst (2018) recommended slow and incremental program
implementation, incorporation of corrections officers and staff into the planning process,
and constant assessment and evaluation of program efficacy.
Serious mental illnesses could contribute to the likelihood that ex-inmates would
re-offend in multiple ways; thus, researchers such as Hirschtritt and Binder (2017) have
contended that the cycle of mental illness-incarceration-recidivism must be broken. Many
individuals who commit crimes, both violent and non-violent, experienced mental illness
(Hirschtritt & Binder, 2017). In many cases, these individuals were incarcerated rather
than being sent to a mental treatment facility. Once in prison, mental healthcare and
resources could be scarce. Trauma and adverse experiences in the prison environment
could further exacerbate mental health conditions, leading to the condition remaining the
same or worsening by the time an inmate was released from prison. Left in a worsened
mental state with few job prospects and financial resources due to imprisonment, these
individuals were then at a significantly heightened risk of reoffending (Hischtritt &
Binder, 2017). Thus, improving mental illness treatment available in prisons and reducing
the number of mentally ill individuals who were incarcerated rather than treated may be
the only way to break the cycle.
Initiatives aimed at improving substance abuse treatment in prisons may have also
contributed to reduced recidivism rates (Haviv & Hasisi, 2019; Ray, Grommon,
Buchanan, Brown, & Watson, 2015). The Access to Recovery (ATR) initiative provided
both support and clinical treatment for inmates who underwent substance abuse treatment
in prison. The Substance Abuse and Mental Health Services Administration (SAMHSA),
a U.S. government organization, funded and managed the initiative. Because the initiative
was implemented at prisons all over the U.S., some institutions implemented the initiative
more effectively than others. Evaluation of the implications of the initiative revealed that
prison programs guided by ATR were more effective when reentry into society and
preventing recidivism were prioritized alongside the prevention of substance abuse (Ray
et al., 2015). Haviv and Hasisi (2019) also studied addiction treatment as an approach to
preventing recidivism. The implications of three different drug rehabilitation programs
offered in Israeli prisons were studied. Upon conducting a comparative analysis, only one
of the three programs that were studied was found to be effective. The effective program
was more comprehensive than the other two and addressed more elements of their health
and well-being. Haviv and Hasisi (2019) concluded by noting that a two-fold strategy was
key for effective drug rehabilitation and prevention of recidivism: programs should be
“based on the promising components of rehabilitation, that is, cognitive behavioral
therapy, therapeutic community, long duration, intensity, and positive social climate,”
and should “succeed in retaining its participants through completion.” (p. 2742).
Education has also been used as an approach to reduce recidivism (Ellison,
Szifris, Horan, & Fox, 2017; McCorkel & DeFina, 2019; Sellers, 2016). Education
provided inmates with human capital and knowledge that could be used to seek
opportunities upon being released from prison. Thus, effective prison education programs
contributed to reducing recidivism and had the potential to increase employment
opportunities upon an inmate's re-entry into society. A meta-analysis of prison education
programs conducted by Ellison et al. (2017) revealed 18 existing papers that described
prison education programs researched in robust detail. Analysis of the 18 included papers
revealed an average recidivism rate of approximately 64%. Prison education programs
were found to reduce recidivism and improve the likelihood of an ex-inmate being hired
upon their release by 24%. Findings from Ellison et al.’s (2017) study revealed a
significant connection between education programs in prison, post-incarceration
employment, and recidivism.
Online learning had been proposed and tested as a means of reducing the costs of
prison education programs intended to reduce recidivism (Sellers, 2016). Many prisons
approached the adoption of prison education programs with trepidation due to the
significant costs associated with adopting an education program. Education staff must
have been hired, classroom textbooks and materials must have been acquired, and prison
leaders/administrators must have participated in program planning and strategizing. Thus,
online education was an appealing option, as resources could be shared and utilized at a
much lower cost and fewer educational staff members were needed. Though measures
were necessary to ensure prisoners did not misuse internet access and computer
privileges, emerging research evidence suggested that adverse incidents were low and
online education was a favorable alternative to face-to-face classroom learning within the
prison context (Sellers, 2016).
In summation of this section, high rates of recidivism, or reoffending following a
criminal conviction, served to indicate ineffective elements were present within a prison
system (Carr et al., 2016; Mitchell, Cochran, Mears, & Bales, 2017). Researching and
understanding recidivism was imperative because most individuals who served time in
prison were not serving life sentences (Carr et al., 2016). The U.S. had one of the highest
rates of recidivism among industrialized nations. Many factors and predictive variables
were tied to recidivism, such as mental health conditions, substance, abuse, experiences
of hardship, lack of access to resources/job opportunities, age, antisocial personality
scales, distress, criminogenic needs, criminality within the family, gender, family rearing,
risk scales, antisocial behavioral history, substance abuse, and social achievement,
stigma, prison, discrimination, and anger problems (Gendreau & Goggin, 2019; Kirk,
Barnes, Hyatt, & Kearley, 2018). Some predictors, such as heavy consumption of alcohol,
only predicted recidivism among men or women. Other predictors, such as past DUI
arrests and prior breaches of a court order, were specific to individuals who had
perpetrated a specific type of offense.
Programs and initiatives were implemented by criminal justice institutions to
prevent recidivism. Programs took place during incarceration, during offenders’ release
period, or over a long-term period to permanently reintegrate ex-inmates into their
communities. Slow and incremental program implementation, incorporation of
corrections officers and staff into the planning process, and constant assessment and
evaluation of program efficacy had been shown to improve the effectiveness of
recidivism prevention programs. Initiatives aimed at improving substance abuse and
mental illness treatment may have also contributed to reduced recidivism rates (Haviv &
Hasisi, 2019). Education had also been used as an approach to reduce recidivism, as there
was a significant connection between education programs in prison, post-incarceration
employment, and recidivism.
Extant recidivism research had largely centered on male experiences, as they were
treated as the “norm” in research because they represented most global prison
populations. However, gender-informed interventions were significantly more effective
than gender-neutral interventions where criminal recidivism was concerned. There
remained a need to increase consideration of female experiences and perspectives when
developing interventions aimed at reducing recidivism.
Summary
In summation of this review of literature, the aim of this study explorec the
relationship between criminal thinking, psychological wellbeing, age, and recidivism
among previously-incarcerated violent female offenders that had been released from
prison in the last two years using Yochelson and Samenow’s criminal thinking theory.
Quantitative methods were used to address the purpose of the study. The theoretical
framework that guided this research was Yochelson and Samenow’s theory on criminal
thinking.
Criminal thinking theory (Yochelson & Samenow, 1976; 1977) was a useful
approach for framing research that considered the causes and nature of criminality
alongside recidivism, mental illness, gender-related differences, and other subtopics
(Mandracchia et al., 2015; Morgan et al., 2015). Criminal thinking primarily affected
thinking patterns, automatic errors of thinking, and the process whereby ideas were
translated into a response or action (Yochelson & Samenow, 1976; 1977). Relationships
between mental health and criminogenic thinking, as well as certain mental illnesses and
criminogenic thinking, had been found in recent research; however, additional research
was needed to understand criminal cognitions over time and whether criminal cognitions
change based on different environmental factors (Mandracchia et al., 2015). High
rates of recidivism were indicative of ineffective prison outcomes to a certain degree
(Carr et al., 2016). Predictors of recidivism that had been discussed frequently in recent
literature include mental health conditions, substance, abuse, experiences of hardship,
lack of access to resources/job opportunities, age, gender, antisocial personality scales,
criminal thinking, distress, criminogenic needs, criminality within the family, gender,
family rearing, risk scales, antisocial behavioral history, substance abuse, and social
achievement, stigma, prison, discrimination, anger problems, and alcohol abuse.
Interventions designed to prevent recidivism were implemented by some criminal justice
institutions during incarceration, during offenders’ release period, or over a long-term
period following their release (Gobeil et al., 2016). Some programs involved substance
abuse or mental illness treatment; others center on education and provided resources to
ease the transition of reentry into society.
Chapter 3 offers details of the methodology selected for this research. The purpose
and research questions were reviewed, followed by the role of the researcher.
Aspects of the methodology were then explained, including the instrumentation and data
analysis plan. Threats to validity were then explained within the context of the study.
Ethical procedures that followed would also be explained. A summary will conclude the
chapter.
Chapter 3: Research Method
Introduction
The specific topic explored in this study was the relationship between criminal
thinking, psychological wellbeing, age, and recidivism among previously-incarcerated
violent female offenders that had been released from prison in the last 2 years using
Yochelson and Samenow’s criminal thinking theory was a predictor of recidivism (Folk
et al., 2016; Pimlott Kubiak et al., 2015; Pantalone et al., 2018). Based on the gap in
literature, additional research was needed to understand the relationship between criminal
thinking and the recidivism of violent female offenders, particularly as mental health and
criminal thinking were both factors related to potential predictors of recidivism
(CollicaCox & Furst, 2018). Given this purpose, the following research questions and
hypotheses were developed to guide the proposed study:
RQ1: Is there a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment?
H01: There is no statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment.
Ha1: There is a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment.
RQ2: Is there a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
H02: There is no relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
Ha2: There is a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
RQ3: Is there a statistically significant association between the age of female
offenders and the number of incarcerations?
H03: There is no statistically significant association between PICTS scores related
to age and number of incarcerations.
Ha3: There is a statistically significant association between PICTS scores related
to age and number of incarcerations.
RQ4: Is there a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment?
H04: There is no relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment.
Ha4: There is a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment.
The remainder of this chapter explores, in detail, the proposed methodology and
procedures for this study. I start with an explanation of the role of the researcher, an
indepth description of the methodology, including participant selection and criteria, and
instrumentation that would be used for this study. Next, I present a section on the
procedures that would be used to recruit participants and collect data, followed by a data
analysis plan. The chapter concludes with ethical procedures and a summary of the
chapter’s most salient points.
Research Design and Rationale
This study used a comparative, nonexperimental, cross-sectional design and a
quantitative methodology. A cross-sectional study, rather than a longitudinal study, was
more fitting for this study since data collection involved survey questionnaires that only
occurs during a single period (Asiamah, Mends-Brew, & Boison, 2019). Additionally,
because I explored the differences between female offenders and nonoffenders at one
specific period, I used a nonexperimental approach because there would be no
manipulation of variables or the random assignment of participants. Finally, this study
was comparative since I utilized test norms to compare their results to the offender group.
Population
The primary target population for this study included the adult female violent
exoffenders in Central Texas. The ideal sample size for this population was 128
participants. The minimum sample size for this study was 82, which was computed using
power analysis. The required number of samples was determined through power analysis.
Power analysis was conducted through G*Power software. The sample size computation
was based on different factors. These included the type of statistical analysis consisting of
Cohen’s effect size, level of significance, the statistical power, and the probability of
rejecting a false null hypothesis.
I used a quantitative method, which was an appropriate choice since the objective
of the study was to measure variables and analyze them using a statistical analysis to
explain the phenomena (Mustafa, 2011). A quantitative research design was also a better
choice given that one objective of this study was to examine the potential relationships
and differences between the identified variables. In order to determine association of the
results to the characteristics of violent female offenders, test norms were used consisting
of female nonoffenders.
Sampling and Sampling Procedures
A computation of the ideal sample size was conducted using the G*power analysis.
First, an a priori power analysis was conducted with the following factors: (a) statistical
test of means: difference between two independent means (two groups), (b) two-tailed
test, (c) medium effect size of 0.50 for an independent sample ANOVA, (d) level of
significance of 0.05, and (e) statistical power of 0.80, which was normally used in
quantitative studies (Faul et al., 2009). The computation yielded a minimum sample size
of 128 samples, 64 samples for each of the two sample groups of (a) recently released
female violent offenders and (b) normative control sample of non-offenders. Second, an a
priori power analysis was conducted with the following factors: (a) statistical test of
correlation: point biserial model, (b) two-tailed test, (c) medium effect size of 0.50 for a
correlation analysis, (d) level of significance of 0.05, and (e) statistical power of 0.80.
The computation yielded a minimum sample size of 82 samples. The higher between the
computations used was 128 samples. This meant that there would ideally be at least 128
samples as the minimum to achieve the required statistical power for a quantitative study
of 80% using both the statistical analyses of independent sample ANOVA test and
correlation analysis. Thus, the target sample size for this study, as based on a power
analysis, was 128 which included a breakdown of at least 64 samples of recently released
female violent offenders (study group) and at least 64 samples of normative control
sample of nonoffenders (control group).
Inclusion criteria for ex-offenders was that they were adult females 18 years of
age or older, had served any length of prison time for violent-based offenses in Texas and
must have currently been released from prison. However, inclusion criteria did not
include specific socioeconomic backgrounds, race, ethnicity, or sexuality. I posted an
online flyer survey around the community and at number of local facilities. These
facilities included sober living facilities, nonprofit organizations, probation officials,
halfway houses, substance abuse treatment programs, and street poles/signs in Central
Texas in order to gain access to the adult female population of violent ex-offenders.
Given the specific criteria needed as well as the difficulty in reaching this specific
population, purposive sampling was used. A short inclusion inquiry presented along with
a document of implied informed consent was used before the instruments for this study in
order to assess that the participants met the inclusion criteria. Any individuals who did
not meet the criteria were included in the study.
For the sample of participants in the test norms, inclusion criteria must have
included that they were adult females over the age of 18, have had never been convicted
of a crime, and lived in Central Texas. This critera ensured that there were parallel
demographics with the ex-offenders.
Procedures for Recruitment, Participation, and Data Collection
Before any preliminary contact with potential participants, I received Institutional
Review Board (IRB) approval and permission to conduct research. After receiving
notification from IRB, I began working on creating my survey and scouting potential
areas to advertise my online survey. Once survey was posted, I discussed both academic
and government policies that pertained to ethical standards, including confidentiality and
the importance of maintaining the anonymity of the participants within the survey.
The Implied Informed Consent Form provided background information regarding
the study, a description of the study questionnaires, the purpose of the study, the
directions for completion in the study, the participant inclusion criteria, a statement
concerning anonymity and voluntary participation, the risks associated with participation,
the website location for research results, and my contact information.
The assessments and surveys were administered online to allow for
standardization and minimized any potential bias associated with in-person
administration. The platform utilized to administer the assessments was Survey Monkey.
Survey Monkey was selected due to feasibility and low costs associated with the
platform. The flyer was continuously posted on social media, and in local agencies. The
data was recorded on Survey Monkey and remained in a password-protected file on my
personal computer that always stayed in my possession throughout the study.
I posted the online flyer survey around the community and at number of local
facilities. These facilities included sober living facilities, nonprofit organizations,
probation officials, halfway houses, substance abuse treatment programs and local areas
around the neighborhoods (light posts, mail boxes, street poles) in Central Texas. I also
posted the flyer on social media. There were no conflicts of interest anticipated with this
study. The informed consent included clear instructions on the procedures of the study.
Once the participant accepted the informed consent and continued to the link of the online
survey, it was implied that they agreed to the terms of the study.
Instrumentation
Three instruments were used for this study. The first was the PICT. The second
was the MPWB. Last, I used a demographic questionnaire.
Psychological Inventory of Criminal Thinking
The first was the PICTS, which was an 80-item self-report measure designed that
aimed to assess crime-supporting cognitive patterns. PICTS did so by measuring eight
thinking patterns that were believed to be associated with a criminal lifestyle (Walters,
1995), based on Walters’ (1990) lifestyle model, which suggested that criminal behavior
was ground in specific lifestyles that could be liked to four specific behavioral styles:
interpersonal intrusiveness, irresponsibility, self-indulgence, and social rule breaking
(Palmer & Hollin, 2003). The first version of the PICTS was created in 1989 and had 32
items, with four items for each thinking style, all of which were rated on a 3-point
Likerttype scale (agree, uncertain, disagree). One year later, the PICTS was revised by
adding two validity scales – confusion and defensiveness – as well as a revision to the
Likert scale to form a 4-point rating scale (strongly agree, agree, uncertain, disagree).
Two years after the second version, the PICTS was revised once again, wherein the
number of items for each scale was doubled – from four up to eight – as well as adding
revised validity scales, factor scales, and content scales. Finally, the fourth version of the
PICTS contains eight new fear-of-change items, bringing the PICTS to its current form:
an 80item inventory composed of two validity scales (revised Confusion scale [Cf-r] and
revised Defensiveness scale [Df-r]), eight thinking-style scales (Mollification scale [Mo],
Cut- off scale [Co], Entitlement scale [En], Power Orientation scale [Po], Sentimentality
scale [Sn], Superoptimism scale [So], Cognitive Indolence scale [Ci], and Discontinuity
scale [Ds]), four factor scales (Problem Avoidance scale [PRB], Interpersonal Hostility
scale [HOS], Self-Assertion/ Deception scale [AST], and Denial of Harm scale [DNH]),
two general content scales (Current Criminal Thinking scale [CUR] and Historical
Criminal Thinking scale [HIS]), and one special scale (Fear of Change scale [FOC]). The
survey had been shown to have well-established internal consistency, test-rest reliability,
and temporal stability of the PICTS scales (Walters, 2002).
Measure of Psychological Well-Being
The second instrument used during this research was the MPWB. The MPWB,
developed by Choi et al. (2014), was a short-form scale designed to assess psychological
well-being among adults of all ages. This seven-item Likert-type scale was intended to
measure seven constructs of well-being on the basis of whether participants agreed not at
all (1), agreed a little (2), or agreed a lot (3). The instrument specifically measured the
constructs of purpose in life, self-acceptance, personal growth, acceptance of living
situation, perceived constraints, personal mastery, and self-efficacy. Items three and five
were reverse-coded.
Demographic Questionnaire
The third instrument used in the study was a simple demographic questionnaire
consisting of approximately 10 questions that were also administered to participants. The
demographic questionnaire covered basic questions pertaining to the demographic
characteristics of participants, including age, race, socioeconomic status, and education
level. While not all of this information may have been pertinent, the collection of
demographic data provided the researcher with a more informed basis to analyze the data.
Data Analysis Plan
I entered the data gathered from the responses of participants in the Survey
Monkey into SPSS v23.0 to prepare for data analysis. The study included an analysis of
the demographic characteristics of participants using descriptive statistics, such as
frequencies and percentages for categorical variables and measures of central tendencies
for continuous variables. I planned to also calculate each participant’s scores for the
domains of the two assessments.
RQ1: Is there a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment?
H01: There is no statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment.
Ha1: There is a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment.
RQ2: Is there a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
H02: There is no relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
Ha2: There is a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
RQ3: Is there a statistically significant association between the age of female
offenders and the number of incarcerations?
H03: There is no statistically significant association between PICTS scores related
to age and number of incarcerations.
Ha3: There is a statistically significant association between PICTS scores related
to age and number of incarcerations.
RQ4: Is there a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment?
H04: There is no relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment.
Ha4: There is a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment.
Analysis Plan
The analysis for (RQ1) used an ANOVA which tested the significance of group
differences between two or more. I conducted linear regression analysis to determine
whether statistically significant correlations existed between the PICTS scores of recently
released female violent offenders and normative control sample of non-offenders.
The analysis for (RQ2) was conducted with a linear regression analysis to
determine if statistically significant correlations existed between the PICTS scores and
age of first incarceration (via the PICTS).
The analysis for (RQ3) was conducted further with a linear regression analysis to
determine if statistically significant correlations existed between the PICTS scores of the
ages of the recently released female violent offenders and their number of incarcerations
(via the PICTS).
The analysis for (RQ4) utilized an ANOVA which tested the significance of group
differences between two or more. I conducted linear regression analysis to determine
whether statistically significant correlations existed between the MPWB scores of
recently released female violent offenders and normative control sample of nonoffenders.
For this study, there were two different statistical analyses that were conducted.
These included an independent sample ANOVA which tested the significance of group
differences between two or more; and a correlation analysis to address research question
three. The ANOVA between two independent groups was conducted to determine
whether there was a statistically significant difference between the criminal thinking
scores and the normative control sample of non-offenders as measured by the PICTS
assessment, whether there was a relationship between age of first incarceration PICTS
score, and whether there was a statistically significant association between the age of
female offenders and the number of incarcerations. The power analysis conducted was for
both statistical analyses.
Threats to Validity
Validity was a crucial component in research as it establishes the ways in which
findings of the study lead to valuable conclusions. While this study’s research
methodology ultimately determined its validity, it was also important to note that the
validity of an instrument was also significant. That was why the instruments being used
for this study had been validated. The PICTS was found to have moderate to moderately
high internal consistency and test-retest stability, and meta-analyses of studies in which
the PICTS has been administered found that the PICTS scales were able to predict future
adjustment/release outcome at a low but statistically significant level (Walters, 2002).
The influence of confounding variables could threaten the external validity of a
study. To combat this issue, a demographic questionnaire was administered to
participants. Gathering demographic information helped to ensure that similarities and
differences between the control and non-control samples were rooted in this study’s
central concepts, rather than differences on the basis of education, age, race, and other
demographic traits (Persaud & Mamdani, 2006). The questionnaire acted to sort
participants and ensure representation in the population being studied, which made the
results more generalizable.
Moreover, for the purpose of this study, I attempted to control internal threats to
validity of maturation by making sure that the study happened within a fixed period. The
participants had four weeks to respond to the initial recruitment letter. Conducting this
study in such a timely manner helped to ensure that data obtained was relevant and able
to provide valid conclusions. Participants were selected purposively and had membership
in the community under investigation, ensuring that data collected was relevant to the
research questions.
Ethical Procedures
Before starting the data collection processes, this research study received IRB
approval from the University. Because this study used a survey method, which involved
humans as participants, I made sure to protect the anonymity and of participants.
Participants’ anonymity stayed protected with an implied informed consent form attached
to the first page of the online survey. Moreover, there were no identifiable information
from the participants and all data remained anonymous.
Each participant seen an informed consent page before taking their survey. To
ensure anonymity of participants, there were no identifiable information, such as name or
address, collected. Only aggregate data appeared in any published work. Participants
were advised that they may leave the study at any time without penalty, and that their
participation would not impact anything in their personal or professional lives.
All the data collected in this study stayed secure in a password-protected computer
and personally kept safely guarded. All surveys and documentation for the current study
remains for five years after the completion of this study, after which it would be deleted.
In addition, there were no foreseeable adverse events triggered by the participation in this
study or by the use of the surveys or assessments for participants, and no conflicts of
interest were anticipated. Participants were informed that their participation was
anonymous and the study was voluntary. They were assured that they could stop at any
time.
Summary
The purpose of the quantitative research study explored the relationship between
criminal thinking, psychological wellbeing, age, and recidivism among
previouslyincarcerated violent female offenders that had been released from prison in the
last two years. The group of interest in this study were female violent offenders in the
region of Central Texas. The ideal sample size for this population was 128 participants.
The minimum sample size for this study was 82, which was computed using power
analysis. In addition, a small number of non-offender participants were also recruited as
a control group. After signing an Informed Consent form, all participants took an online
flyer survey, the Psychological Inventory of Criminal Thinking Styles (PICTS), and the
Measure of Psychological Well-Being (MPWB). Data analysis included determining
associations between criminal thinking in ex-offenders and non-offenders, between
criminal thinking for both ex-offenders and non-offenders, as well as an association with
the PICTS scores. Chapter Four will present the results of this study.
Chapter 4: Results
Introduction
The purpose of this quantitative research study was to analyze the relationship
between criminal thinking, age, psychological well-being, and recidivism among recently
released female violent offenders through the use of Yochelsen and Samenow’s criminal
thinking theory. The following research questions and hypotheses were addressed:
RQ1: Is there a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment?
H01: There is no statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment.
Ha1: There is a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment.
RQ2: Is there a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
H02: There is no relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
Ha2: There is a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
RQ3: Is there a statistically significant association between the age of female
offenders and the number of incarcerations?
H03: There is no statistically significant association between PICTS scores related
to age and number of incarcerations.
Ha3: There is a statistically significant association between PICTS scores related
to age and number of incarcerations.
RQ4: Is there a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment?
H04: There is no relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment.
Ha4: There is a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment.
The following is a discussion of the study’s population and sample as well as a
demographic description of the sample. Demographic descriptions included frequencies
and percentages for categorical (nominal) variables and descriptive statistics of minimum,
maximum, mean, and standard deviation for variables measured at the interval level of
measurement. Also presented were the testing of parametric assumptions for the
statistical analysis and results of hypothesis testing. This chapter concludes with a
discussion of the results of this study.
The second research question's findings conform to the theory on criminal
thinking, there were more offenders with only a high school education or lower in the
offender group compared to the nonoffenders group. This shows lower levels of
education and lack of access to education may have impacted incarceration rate,
recidivism, and criminal thinking (Mandracchia et al., 2015). Consequently, their
thinking was decidedly different from ordinary individuals who pick up these lessons
over time. Criminal thinking theory was a useful approach for framing research that
considers the causes and nature of criminality alongside recidivism, mental illness and
other subtopics (Mandracchia et al., 2015; Morgan et al., 2015). The findings of this
study support the theory on criminal thinking (Yochelson & Samenow, 1976; 1977) and
further demonstrates distinctions between general society and the criminal population.
Recidivism was more likely to take place in individuals who had higher
frequencies of crimes. Relationships between mental health and criminogenic thinking
and certain mental illnesses and criminogenic thinking had been found in recent research;
however, additional research was needed to understand criminal cognition over time and
whether criminal cognition changes based on different environmental factors
(Mandracchia et al., 2015). For example, an individual who was mentally preoccupied
with committing a crime would likely do so which would increase incarceration risk. If
he did not reform his thinking patterns in prison, there was a higher risk of recidivism
once he was released because this is the familiar choice for him (Samenow, 2014).
The results of the fourth research question on the association of mental well-being
and criminal thinking indicate that well-being was an independent factor that can regulate
criminal thinking (Walters, 2107). Criminal thinking theory did not explore causative
factors affecting criminal thinking but acknowledged that criminals have distinctly
different thinking patterns (van Ginneken, 2015; Vrabel et al., 2019). Hence, as far as
well-being was concerned, the theory worked beyond its purview.
Data Collection
The primary target population for this study included the adult female violent
exoffenders in Central Texas. Inclusion criteria for ex-offenders were that they were adult
females 18 years of age or older, had served any length of prison time for violent-based
offenses in Texas, and were released from prison at the time of the study. In addition,
nonoffender participants were also recruited as a control group. The platform utilized to
administer the assessments was Survey Monkey. Survey Monkey was selected due to
feasibility and low costs associated with the platform. A flyer which explained the
purpose of the study was continuously posted on social media and around the community
of Central Texas. A link to the survey was provided on the advertised flyer which
directed them to the assessments which included a demographic survey, the PICTS
survey, and the MPWB survey.
Demographics
There were a total of N = 98 female participants in the study of which 70 (71.4%)
were ex-offenders and 28 (28.6%) were not ex-offenders (Table 1).
Table 1
Are you a Female Ex-Offender?
Frequency Percent
No 28 28.6
Yes 70 71.4
Total 98 100.0
Table 2 provides the distribution of age categories of the ex-offenders and
nonoffenders. Within the nonoffender group, most participants were in the 25-34
category, 11(39.3%), whereas ex-offenders were mostly in the 35-44 age category,
23(32.9%). In both groups, there were few people in the 55-64 and 65+ age categories.
Among ex-offenders, there were 4(5.7%) in the 55-64 group and 2(2.9%) in the 65+ age
category. Among nonoffenders, there were 6(21.4%) in the 55-64 group and 1(3.6%) in
the 65+ group.
Table 2
Age Categorization
Are you a female ex-offender? Frequency Percent
No
25-34
35-44
45-54
55-64
11
8
2
6
39.3
28.6 7.1
21.4
65+ 1 3.6
Total
28
100.0
Yes
18-24
25-34
35-44
45-54
13
16
23
12
18.6 22.9
32.9
17.1
55-64 4 5.7
65+ 2 2.9
Total 70 100.0
In order to determine if the distribution of ages were associated with the type of group
(ex-offender or nonoffender) the Chi-Square test for association was conducted. The
ChiSquare test of association was used to determine the level of association between two
nominal variables. Table 3 provides the results of the Chi-square test which indicate that
there was a significant association between group type (ex-offenders versus
nonoffenders) and age category, χ 2(5) = 13.986, p = .010. The Fisher’s Exact test was
used since there were five cells that had an expected count less than five.
Table 3
Chi-Square Tests for Age
Value df Asymp. Sig.
(2sided)
Exact Sig.
(2sided)
Pearson Chi-Square 13.549a 5 .019 .016
Fisher's Exact Test 13.986 .010
N of Valid Cases 98
Note. 5 cells (41.7%) have expected count less than 5. The minimum expected count was
.86.
Table 4 provides a breakdown of the distribution of race by ex-offenders and
nonoffenders. The proportions of Black or African Americans were similar in both
groups: Nonoffenders 10 (35.7%) and ex-offenders 28(40.0%). There was a large
discrepancy in the proportions of White women between nonoffenders, 15(53.6%) versus
ex-offenders,
22 (31.4%).
Table 4
Race
Are you a female ex-offender? Frequency Percent
No
Black or African American
Hispanic
White
10 3
15
35.7
10.7
53.6
Total
28
100.0
Yes
Asian / Pacific Islander
Black or African American
Hispanic
3
28
17
4.3
40.0
24.3
White 22 31.4
Total 70 100.0
In order to determine if the distribution of races were associated with the type of group
(ex-offender or non-offender) the Chi-Square test for association was conducted. Table 5
provides the results of the Chi-square test which indicate that there was no significant
association between group type (ex-offenders versus non-offenders) and race, χ 2(3) =
4.971, p = .147. The Fisher’s Exact test was used since there were two cells that had an
expected count less than five.
Table 5
Chi-Square Tests for Race
Value df Asymp. Sig.
(2sided)
Exact Sig.
(2sided)
Pearson Chi-Square 5.697a 3 .127 .125
Fisher's Exact Test 4.971 .147
N of Valid Cases 98
Note. 2 cells (25.0%) have expected count less than 5. The minimum expected count
was .86.
Table 6 provided the distribution of educational level of ex-offenders and
nonoffenders. There was a considerable amount of individuals with lower educational
levels (less than high school or high school) in the ex-offender group compared with the
nonoffenders. Within the nonoffender group, 29 (41.4%) had less than a high school
education and 30 (42.9%) had only a high school education. There were less
nonoffenders with lower educational levels: 1 (3.6%) less than high school and 8 (28.6%)
high school education. Also, there were very few people with higher educational levels
among ex-offenders: 1 (1.4%) with a Bachelors degree and 9 (12.9%) with some college.
In the nonoffender group, 5 (17.9%) had some college, 7 (25.0%) had a Bachelor’s
degree, and 7 (25.0%) acquired a graduate degree.
Table 6
Education
Are you a female ex-offender? Frequency Percent
No
<HS
HS/GED
Some college
Bachelors
1
8
5
7
3.6
28.6 17.9
25.0
Graduate degree 7 25.0
Total
28
100.0
Yes
<HS
HS/GED
Some college
Bachelors
29
30
9
1
41.4 42.9
12.9
1.4
Missing 1 1.4
Total 69 98.6
In order to determine if the distribution of educational levels were associated with the
type of group (ex-offender or nonoffender) the Chi-Square test for association was
conducted. Table 7 provides the results of the Chi-square test which indicate that there
was a significant association between group type (ex-offenders versus nonoffenders) and
education level, χ 2(4) = 40.024, p < .001. The Fisher’s Exact test was used since there
were four cells that had an expected count less than five.
Table 7
Chi-Square Tests for Education Level
Value df Asymp. Sig. (2-sided) Exact Sig. (2-sided)
Pearson Chi-Square 41.619a 4 .000 .000
Fisher's Exact Test 40.024 .000
N of Valid Cases 97
The Psychological Inventory of Criminal Thinking Styles (PICTS) Survey
As mentioned earlier, the Psychological Inventory of Criminal Thinking Styles
(PICTS) survey was used in order to measure individuals’ level of criminal thinking
styles. The PICTS is an 80-item self-report measure designed that aims to assess
crimesupporting cognitive patterns. The items were measured on a Likert scale to form
a 4point rating scale (disagree, uncertain, agree, strongly agree). Reliability was
measured by calculating Cronbach’s alpha. A general accepted rule was that α of 0.6-
0.7 indicates an acceptable level of reliability, and 0.8 or greater a very good level.
Nunnally (1978) recommends a minimum level of .7. Reliability was calculated as .982
which indicates excellent reliability for the PICTS. As a result, the mean of item
responses was calculated and served as a measure of PICTS used in the analysis.
The Measure of Psychological Well-Being (MPWB) Survey
Measure of Psychological Well-Being (MPWB) Survey was used in order to
measure overall well-being. The MPWB, developed by Choi et al. (2014), is a short-form
scale designed to assess psychological well-being among adults of all ages. The responses
ranged from 1 = all of the time to 5 = none of the time. Reliability was calculated with a
Cronbach’s alpha of 0.834 which was very good reliability. The mean of item responses
was calculated and served as a measure of well-being used in the analysis.
Wellbeing ranged from 1.00 to 4.60 (M = 2.43, SD = 1.06) and PICTS ranged
from 1.21 to 3.88 (M = 2.90, SD = .98). This information was provided in Table 8.
Table 8
Descriptive Statistics of Wellbeing and PICTS
N M SD
Wellbeing 95 1.00 4.60 2.43 1.06 .711 -.826
PICTS 95 1.21 3.88 2.90 .98 -.875 -1.051
Parametric Testing of Assumptions
The assumptions of normality and absence of outliers were first tested. Skewness
and kurtosis index were used to identify the normality of the data. The results suggested
the deviation of data from normality was not severe as the value of skewness and kurtosis
index were below 3 and 10 respectively (Kline, 2011). Hair et al. (2010) and Bryne
(2010) argued that data was considered to be normal if skewness was between ‐2 to +2
and kurtosis was between ‐7 to +7. Table 9 provides the ranges of standardized values for
wellbeing and PICTS. There were no standardized values beyond -3/+3, thus there were
no outliers in the dataset.
Table 9
Ranges of Standardized Scores
N Min Max
Wellbeing 95 -1.35 2.04
PICTS 95 -1.73 .99
The assumption of equality of variances was tested in order to determine if the
variances of PICTS were similar between ex-offenders and non-offenders. A significant
Min Max Skewness Kurtosis
Levene’s test indicated that there was a violation of this assumption, p = .042. Therefore,
a Welsh’s t test was used which compensates for this violation.
Lastly, linearity was tested in order to determine if there was an approximate liner
relationship between PICTS and wellbeing scores. The scatter plot in Figure 1 below
indicates an approximate negative linear relationship between PICTS and wellbeing
scores. An increase in an individual’s wellbeing seems to be associated with a decrease in
criminal thinking.
Figure 1
Scatter Plot Depicting the Negative Relationship Between Wellbeing and PICTS
Results
An independent t-test was conducted in order to address this first research
question and hypothesis:
RQ1: Is there a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment?
H01: There is no statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment.
Ha1: There is a statistically significant difference between the criminal thinking
scores of recently released female violent offenders and the normative control sample of
non-offenders as measured by the Psychological Inventory of Criminal Thinking Styles
(PICTS) assessment.
There were 26 non-offenders and 20 ex-offenders. An independent-samples t-test
was run to determine if there were differences in the criminal thinking scores (PICTS) of
non-offenders and ex-offenders. There were no outliers in the data, as assessed by
standardized values. PICTS scores were normally distributed, as assessed by skewness
and kurtosis indexes but the assumption of homogeneity of variances was violated, as
assessed by Levene's test for equality of variances (p = .042). The mean criminal thinking
scores were greater in the ex-offender group (M = 3.47, SD = 0.30) than non-offenders
(M = 1.38, SD = 0.17), a statistically significant difference, M = 2.09, 95% CI [1.99,
2.19], t(80.586) = -42.787, p < .001. Tables 10 and 11 provide this information.
Table 10
PICTS Score
Are you a female ex-offender? N Mean Std. Deviation Std. Error Mean
No
PICTS
Yes
26
69
1.38
3.47
.17
.30
.03
.04
Table 11
Independent t-Test (Equal Variances Not Assumed)
t df p
Mean
Difference
Std. Error
Difference
95% CI of the
Difference
Lower Upper
-42.787 80.586 .000 -2.09 .05 -2.19 -1.99
A one-way ANOVA was conducted in order to address this second research
question and hypotheses:
RQ2: Is there a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
H02: There is no relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
Ha2: There is a relationship between the age of first incarceration and the score
measured by the Psychological Inventory of Criminal Thinking Styles (PICTS)
assessment.
A one-way ANOVA was conducted to determine if PICTS scores of ex-offenders
were different among age categories. There were no outliers in the data, as assessed by
standardized values. PICTS scores were normally distributed, as assessed by skewness
and kurtosis indexes and the assumption of homogeneity of variances was met, as
assessed by Levene's test for equality of variances (p = .437). Participants were classified
into six age groups: 18-24 (n = 13), 25-34 (n = 15), 35-44 (n = 23), 45-54 (n = 12), and
65+ (n = 2). PICTS score was greatest for the 55-64 age group (M = 3.57, SD = 0.20) and
the lowest score was in the 25-34 age group (M = 3.38, SD = 0.47). The differences in
mean PICTS between the age categories, was not statistically significant, F(5, 63) =
0.430, p = .826. Tables 12, 13, and 14 provide this information.
Table 12
Descriptive Statistics of PICTS by Age
Age Mean Std. Deviation N
18-24 3.45 .32 13
25-34 3.38 .47 15
35-44 3.50 .23 23
45-54 3.51 .20 12
55-64 3.57 .20 4
65+ 3.51 .19 2
Total 3.47 .30 69
Table 13
Levene's Test of Equality of Error Variances
F df1 df2 Sig.
.981 5 63 .437
Table 14
Tests of Between-Subjects Effects
Source Type III Sum of
Squares
df Mean Square F p
Corrected Model .206 5 .041 .430 .826
Intercept 429.366 1 429.366 4477.310 .000
Age .206 5 .041 .430 .826
Error 6.042 63 .096
Total 838.336 69
Corrected Total 6.248 68
The third research question pertained to the relationship between age of female
ex-offenders and the number of incarcerations:
RQ3: Is there a statistically significant association between the age of female
offenders and the number of incarcerations?
H03: There is no statistically significant association between PICTS scores related
to age and number of incarcerations.
Ha3: There is a statistically significant association between PICTS scores related
to age and number of incarcerations.
Information regarding the number of incarcerations was not collected due to
concerns of participants being poor historians with the inability to accurately recall
number of incarcerations thus this could not be directly measured. However, as addressed
in Research question 2, PICTS scores were not significantly different based on age
categories. Therefore, it seems likely that there was no significant association between
age and the number of incarcerations.
Linear regression was used in order to address this fourth research question and
hypothesis:
RQ4: Is there a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment?
H04: There is no relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking as measured by the
Psychological Inventory of Criminal Thinking Styles (PICTS) assessment.
Ha4: There is a relationship between the mental well-being as measured by the
Measure of Psychological Well-Being (MPWB) and criminal thinking
To assess linearity a scatterplot of wellbeing against PICTS score was plotted.
Visual inspection indicated a linear relationship between the variables. There was
homoscedasticity and normality of the residuals and there were no outliers outside -3/+3
standard deviations. Wellbeing score significantly predicted PICTS score, F(1, 94) =
495.621, p < .001, accounting for 84.2% of the variation in PICTS score (R2 = .842). A
one unit increase in wellbeing leads to a 0.844 decrease in PICTS score. Tables 15, 16,
and 17 provide this information.
Table 15
Model Summaryb
R R Square Adjusted R Square Std. Error of the Estimate Durbin-Watson
.918a .842 .840 .39007 1.763
Note. a. Predictors: (
Table 16
ANOVAa
Constant), Wellbeing; b. Dependent Variable: PICTS.
Sum of Squares df Mean Square F p
Regression 75.410 1 75.410 495.621 .000b
Residual 14.150 93 .152
Total 89.560 94
Note. a. Dependent Variable: PICTS; b. Predictors: (Constant), Wellbeing.
Table 17
Coefficientsa
Unstandardized Coefficients Standardized Coefficients
B Std. Error Beta
t p
(Constant) 4.955 .101 49.256 .000
Wellbeing -.844 .038 -.918 -22.263 .000
Note. a. Dependent Variable: PICTS.
Summary
The purpose of this quantitative research study was to analyze the relationship
between criminal thinking, age, psychological well-being, and recidivism among recently
released female violent offenders through the use of Yochelsen and Samenow’s criminal
thinking theory. Regarding the first research question, there was a significant mean
difference in PICTS scores between ex-offenders and nonoffenders. Mean PICT scores of
ex-offenders were significantly greater than non-offenders. Pertaining to the second
research question, there were no significant mean differences in PICTS scores based in
age categories. No data was collected on the number of incarcerations; thus the third
research question was not addressed. Lastly, pertaining to the fourth research question,
there was a significant negative relationship between wellbeing and PICTS scores.
Increasing wellbeing results in a significant decrease in criminal thinking, as measured by
PICTS score.
What follows in Chapter 5 was a discussion as to how the results of this study
were interpreted in the context of the theoretical framework. Limitations of the results of
the study are provided. Additionally, recommendations for future research are discussed.
Chapter 5: Discussion, Conclusions, and Recommendations
The purpose of this quantitative research study was to analyze the relationship
between criminal thinking, age, psychological well-being, and recidivism among recently
released female violent offenders through the use of Yochelsen and Samenow’s criminal
thinking theory. This study contributed to the understanding of how criminal thinking
influences recidivism of formerly incarcerated female violent offenders (Yochelson &
Samenow, 1976; 1977).
This section was aimed at analyzing the findings of the study and focusing on
ways those findings can promote positive social change. The first research question
addressed in this study was whether offenders were more likely to have criminal thinking,
the study's findings were affirmative. There was a statistically significant difference
between criminal thinking scores of recently released female violent offenders and
nonoffenders. The second research question was whether offenders' age at first
incarceration was likely to influence criminal thinking, there was no relationship between
the age of the first incarceration and levels of criminal thinking. The third research
question was whether offenders' age influenced the number of incarcerations, however
the number of incarcerations was not collected due to concerns of participants being
inaccurate. The fourth research question was whether offenders' well-being influenced
criminal thinking, the study's findings were affirmative. There was a relationship between
mental well-being and criminal thinking.
Interpretation of the Findings
In this study demographic data was also collected. Age of participants and
educational levels were included in the demographic data. The following two findings
speak to this set of demographics.
•There were more younger female offenders (under 45 years) compared to the
older female offenders who participated in this study.
•Offenders were more likely to have dropped out of high school or only had a
high school degree than the non-offender group.
These findings were consistent with previous studies that have demonstrated that low
education levels often lead to dire economic conditions such as poverty, which also
triggered criminal activity (Machin, Marie & Vujić, 2011; Kearney et al., 2014). Hence a
person with a low level of education was likely to have fewer options of earning an
income and may resort to crime. This phenomenon was also evident in low-income
neighborhoods with lower mean levels of education and higher crime rates. Thus, this
study reinforced previous scholars' findings who had ably demonstrated the connection
between poor education and crime.
The next pertinent finding is the role of criminal thinking.
•There is a statistically significant difference between criminal thinking scores
of recently released female violent offenders and nonoffenders.
Recidivism was more likely to occur among offenders than non-offenders (Mulder et al.,
2011). Studies indicated that serving time may have had the opposite effect on an
offender, hardening them to criminal activity instead of reforming them to become better
society members (Haney, 2012). Therefore, the study's findings were similar to other
studies that had proven this over time, validating previous researchers' claims.
A third interesting finding was the role of age and the theoretical construct of
criminal thinking.
•There was no relationship between the age of the first incarceration and
criminal thinking.
Van der Geest et al. (2016) demonstrated that the first incarceration and criminal thinking
age were significantly related. The lower the age the person began criminal behavior the
more the person had a pattern of criminal thinking. He attributed this phenomenon to the
simple fact that younger inmates get exposed to older, more hardened criminals within
the prison system, which could worsen rather than improve their outlook on crime. Once
such offenders were released from prison, not only were their attitudes more
accommodating of criminal thought and crime, they may have built a network of enablers
in the outside world based on recommendations from prison. These factors could lead an
increase in criminal behavior and increased levels of recidivism. Recidivism had also
been linked to criminal thinking, which involved two central processes within an
offenders' thoughts, reactive and proactive criminal thinking. The former involved
reactions, indicating weak control over impulses and emotions, while the latter suggests
the ability to plan and be deliberate (Walters, 2107). The findings of this study differed
from these findings, which may have suggested that this contradiction was only specific
to the sample group that the researcher examined. Samenow (2014) suggested that all
criminal behavior was a matter of an individual consciously making the wrong choice
regardless of how good or bad their personal circumstances happen to be. Hence in this
context, recidivism would not be due to criminal thinking patterns but a deliberate effort
by the criminal to commit crime.
These findings were consistent with Farrington et al. (2013), who demonstrated
the same factor in their study on the association between age and offenses committed.
However, he pointed out that this fact was only contradicted when considering the case of
professional criminals or gangsters such as members of the mafia, who can repeatedly
serve sentences over their criminal career. Habitual criminals could be jailed several
times over decades, which means that if the sample were focused on specific types of
criminals, then there would be a statistically significant relationship between age and the
number of incarcerations. In such a scenario, the higher the age, the larger the number of
incarcerations the sample may have indicated. Hence, this study was partially true and in
conformity with previous studies.
Fourth, data revealed the role of mental health and criminal thinking.
•There is a negative relationship between mental well-being and criminal
thinking.
Maschi, Viola & Morgen (2014) investigated the associations between mental stability
and behavior. He found that the more stable the study subjects were, the less erratic their
behavior became. This stability was reflected in improved social interactions within their
workplace and home life. Subjects suffering from depression displayed a tendency for
unreliability, lack of discipline, and a poor work ethic. His findings mirror this study's
findings and validate them despite the two research works' key distinctions of stable
behavior and unstable behavior. This study found a significant association between age
distribution and offender type, as demonstrated in a chi-square test. There were younger
female offenders (under 45 years) compared to the older.
The study also found a significant association between education and likelihood of
offending, with the less educated being more likely to commit the crime. Hence, more
offenders with a less than high school or high school only education level in the offender
group than the non-offenders. This also suggested possible future studies for scholars.
A linearity test found an approximate negative relationship between well-being
and likelihood to become an offender. The higher the individual's well-being scores, the
less likely they were to end up offending.
Limitations of the Study
This study had several limitations. First, the fundamental research was
quantitative, gathering data from the respondents using questionnaires with closed-ended
questions. This approach did not allow input from other influential stakeholders such as
prison administrators, psychologists, prison guards and other relevant staff. A suggestion
could be future researchers broadening their participants to include other people in the
prison. Such participants may have introduced hidden but influential aspects of the prison
system, which could have changed the recommendations.
Second, the study was only focused on a small sample of respondents who were
all female. Generalizing this to males or both federal and state prison offenders was not
possible due to the sample's unique nature. Accommodating the possibility of projection
to these larger populations was limited, which would have been possible with a more
heterogeneous sample. Therefore, the study's generalizability and utility were limited to
populations of offenders who possess the same characteristics as those included in this
study.
Significance of the Study
This study had highlighted several factors that affect offenders and how these
translated into their lives outside prison. This research provided a new body of findings
relevant to mental health research about both offenders and non-offenders. Mental health
scholars, judicial stakeholders, policymakers, and others may want to consider how the
findings of this study may influence how release programs and oversight policies were
developed. Furthermore, non-offenders were also likely to be part of the puzzle when
making decisions about how education influences criminal behavior.
This study had also highlighted the importance of increasing education levels to
reduce criminal thinking among likely offenders. The higher the educational level of
individuals, the less likely they were to commit crimes. The more educated individuals
were, the more likely they were to get good jobs, reduce criminal behavior, and reduce
criminal thinking. The more educated individuals were, the more they understand the
consequences of crime and how adversely it would affect their lives. The more educated
individuals were, the higher the likelihood that they would come up with innovative
solutions to life's challenges, eliminating the need for criminal behavior and thinking.
The findings of this study also highlighted the importance of reengineering the
prison system to make it more useful for offenders and society as a whole. Prison reform
activists would therefore find in the study some insights that may further enrich their calls
for the reorientation of the prison system resulting in a more humane system that
prioritizes mental well-being.
The researcher's theoretical framework for this study was Yochelson and
Samenow's theory on criminal thinking, which posited that criminals think differently and
had different personalities than non-criminals because of their mental attitudes. Criminal
thinking and choices theory was a valuable and practical theory for lending insight into
the causes of criminal behavior, recidivism, gender-related differences among
incarcerated individuals, and other research aimed at explaining/examining criminality
(Yochelson & Samenow, 1976; 1977). The two researchers demonstrated that criminals
use distinctly different thought processes than normal individuals, frequently mirroring
doing the opposite of what was expected from an average person, e.g., a typical
individual would generally dislike walking through a crowded street because he would
have to bump into many people. Still, a pickpocket would love the opportunity because
he gets a chance to steal from the unsuspecting public. Their view of the situation was
peculiar and may or may not have justifications, such as blaming the victim for tempting
them into the crime, hence rationalizing the crime they want to commit.
Finally, this study highlighted the importance of reengaging offenders after
release to glean insights from them. A released offender was one of the best windows into
the prison world, and their feedback has enriched this research in a significant way. Using
offenders to formulate prison policy was a reliable approach to reforming the system
because they would expose all its weaknesses with internal perspectives and solutions as
happens in Sweden (Nilsson, 2013). Mental health, education, age, and criminal thinking
should all be considered when developing release programs and oversight.
Recommendations
Researchers may want to consider using a longitudinal study that considers the
variables influencing criminal behavior over time. For example, tracking the recidivism
rates over 20 years to how changes in policy may influence criminal behavior upon
release. Tracking people who did reoffend over 30 years would also provide a deeper
perspective, such as what motivated them to commit crimes repeatedly. The following
will contribute to effective ways towards positive social change.
Prisons should consider prioritizing offenders' mental reorientation rather than
emphasizing confinement and punishment for offenses. The mental well-being among
offenders should be addressed and included in their release plan to assist in reducing
recidivism rates. Ensuring that each offender was engaged in an activity that promotes
their mental well-being in a permanent manner beyond the prison walls would likely
translate into reduced criminal thinking and less recidivism (Mandracchia et al., 2015).
Prison programs need to be extended beyond the prison sentence to ensure the
offenders continues to get support in the crucial first two years once they leave prison.
Activities like counseling, community service, and mentoring the young, would likely
keep an offender engaged in productive activities that open up new opportunities and
reduce criminal thinking. Probation should be a period when the judicial system increases
its engagement with the offenders rather than repeatedly reminding them that they were
under watch and could land in prison at any time.
Future researchers may want to incorporate a mixed-method research design that
would enable them to collect quantitative and qualitative data. Such a design would
accord the study a more encompassing and insightful findings and conclusions. A
longitudinal, mixed-method approach would provide even more profound and richer
findings, illustrating the interplay between government budgeting, prison reforms, and
administrative will. Including a focus group discussion for the prisoners in the qualitative
aspect would enrich the study further and possibly provide revolutionary findings because
the offenders would provide their perspectives, revealing dominant considerations that
may not be readily apparent.
A broader sample of the country's prison population would also ensure future
studies were more generalizable. A sample that considers various strata such as
demographics, prison concentration, population, and crime prevalence would provide
more generalizable findings applicable statewide or nationally.
Implications
Various implications proceeded from the findings and discussions above. First, the
prison institution may not be reforming offenders with a focus on confinement and
punishment for offenses rather than the mental reorientation of convicts. From the
findings, it was evident that recidivist tendencies were more pronounced among offenders
than the general public due to a preponderance of criminal thinking in their thought
patterns. Gemeda (2017) explained that severe crimes and recidivism might be directly
related to psychopathic personality traits, i.e., extreme criminals who eventually served
time and were released were likely to end back in prison because of personality
aberrations. Furthermore, psychopathic personality traits like violence and impulsiveness
also explain the effect of antisocial behaviors, such as crime, on mental health and its
relation to recidivism. These weaknesses were worsened by other complicating factors
such as age, number of sentences, and mental well-being, that were important in an
offender's life. Researchers have suggested that women in prison report emotional and
mental health problems related to the prison environment such as depression and anxiety
(Caulfield, 2016; van Ginneken, 2015). Therefore, the findings illustrate a failure in the
judicial systems that proclaim specific aims but fall far short of them or achieve entirely
different results based on the recidivism rates seen above that show higher crime rates
among ex-convicts compared to non-convicts.
The prisons' focus was to rehabilitate, however it may instead impact an
individual’s depression and criminal thinking (Adams et al., 2017; Crewe et al., 2017;
Walters & Lowenkamp, 2016). The negativity in offenders' thought patterns, highlighting
the bias towards crime among individual offenders who then relapse into recidivism.
Therefore, prison could be considered a potential training ground for criminals rather than
a place where offenders ponder their actions and seek ways to reform. The perpetuation
of more crimes seems easier for offenders than non-offenders, which suggests potential
weaknesses within the prison system's framework. Sardhamar & Telle (2012) found that
recidivism rates dropped 20% in Norway when the focus of prison moved from
retribution to wholesome rehabilitation. In essence, considering the results above, prison
could worsen offenders' mental condition rather than improve and reshape their outlooks
to make them better members of society.
Conclusion
In conclusion, this study set out to answer three research questions regarding the
impact of various factors on criminal thinking and how these relate to recidivism
(Yochelson & Samenow, 1976; 1977). The study indicates that mental health, age, and
criminal thinking may affect recidivism. Using a cross-sectional quantitative design for
data collection, prisons currently constituted may not be very effective in curbing
recidivism. Therefore, the recommendation was that prison stakeholders can apply to
salvage their institutions and minimize the perpetuation of crime in a setting designed to
eliminate crime.
