Correlates Of Recidivism Among Released Prisoners: A Study of Kakamega County, Kenya

Released prisoners in Kenya have a 75% likelihood of committing another crime and a 50% probability of going to jail two years after their discharge from prison custody. From the trend of recidivism in Kenya, there are a staggeringly high number of offenders being incarcerated and eventually released back to the community, and the high risk of re-arrest and re-incarceration is a concern for policymakers, criminologists and correctional managers. This study examined the influence of offender characteristics, offender reintegration and community perception and attitude regarding recidivism in Kakamega County, Kenya. The study adopted a survey research design. Findings reveal a statistically significant relationship between offender characteristics and recidivism. In addition, offender reintegration and community perception and attitude towards offenders greatly influence recidivism. From the study, it is recommended that the government provide correctional officers with the required resources to use the actuarial risk assessment model. The model is applied to the released offenders to predict the future probability of recidivism. In addition, it is recommended that the government and the various correctional stakeholders come up with an integrated approach that specifically targets successful re-entry of offenders upon release from prison. Finally, it is recommended that the government develop programmes targeting awareness of the community members to desist from stigmatising ex-offenders. & Yang, 2005: 186). Correlations in relapsed offenders' recidivism and criminal activity are the product of a complex phenomenon. Strain theories and labeling theory are used in the study. One model, positive psychology is also used. These theories and model explain the relationship between socio-economic, human and environmental factors and recidivism.


The research problem
The Kenya Prisons Service is mandated by the Prisons Act Chapter 90 Laws of Kenya (Republic of Kenya, 2012: 8), to handle rehabilitation and transformation of prisoners by learning, counseling education and career programmes. One of the core functions of the Kenya Prisons Service is rehabilitation and reformation of prisoners for social re-integration. The Probation Service in Kenya is charged with the reintegration and resettlement of offenders released from prisons through the aftercare services provided under the Probation of Offenders Act 11 of 2017.
Despite the resources spent by the State Department of Corrections, which comprises of the Kenya Prisons Service and the Probation Service to rehabilitate, reform and reintegrate offenders, a high rate of recidivism has been recorded in the country. According to the Kenya National Bureau of Statistics Report (Republic of Kenya, 2019: 277) in the year 2018, there were 16 987 recidivists in prisons out of 53 765 average daily prison population representing 35,59 percent. This cyclic movement of offenders to prisons from the community and back to prisons after release will without any doubt contribute enormously to the growth of the prison population and strain the allocated resources. In addition, there will be increased crime rates in society.
Little has been documented about an integrated offender management process in Kenya. The transitional challenges offenders face upon release from prisons have not been adequately addressed. Thus, it becomes imperative through sound research to find out the correlates of recidivism among released prisoners by specifically establishing the role of prisoner reintegration on recidivism, the relationship between offender characteristics and recidivism, and the influence of the community perception and attitude on recidivism. This will adequately address the transitional challenges that offenders face upon release from prisons that influence their re-offending behaviour in addition to developing an integrated approach towards the offender management process.

Research aim and objective
Examine the relationship between offender characteristics and recidivism in Kakamega County 1.4 Research methodology The study on "the correlates of recidivism among released offenders in Kakamega County" has been conducted by survey research design. This design is usually based upon samples whereby instead of directly studying whole populations, surveys typically collect evidence from a small sample of people selected from the population (Jupp, 2002:34). This design aids the researcher in collecting original data to describe a population that is too large to observe directly. some degree of success, these services have limitations in terms of ability and regional scope (Solomon et al., 2004: 309).

Housing and recidivism
In the United States of America, the bulk of discharged prisoners reside near their family members. Approximately three-quarters of Chicago released inmates expected to stay with the family in one analysis of the Urban Institute, and an even greater 88 percent were staying with the family 4 to 8 months later (La Vigne et al., 2004: 39). This is not always an obvious or possible choice, as family members may have been victimised, or otherwise harmed by a returning person. Women offenders experience high rates of abuse and victimization, regularly at the hands of household members, both in childhood and adulthood (Chesney-Lind, 2002: 88;Harlow, 1999: 341;Richie, 2001: 380).
Ex-offenders who have been accused of sexual offenses face additional limits on housing safety. In the United States of America, 47 States and the Federal Government are enforcing Megan's Amendment law which allows people guilty of sexual offenses to enroll.
1.5.1.6 Parenthood, social relationships, and recidivism In the United States of America, State and Federal jails, well under half of the inmates have minor children; twothirds of the women and half of the men in prison were parents (Glaze & Maruschak, 2008: 126). Research by Glaze and Maruschak (2008: 117) found that the child's mother was 88 percent the primary caregiver of the offspring of the incarcerated parents. On the other side, just 37 percent of those with imprisoned moms stay with their fathers; 45 percent of those children are most probable to live with their grandparents and 23 percent with other families. Knowing the peer system and the probability of relapse through criminogenic social networks, especially between male offenders, is the subject of much criminological study (Scott, 2004: 342;Warr, 1998: 204). Those ex-offenders who revive behavioural patterns from pre-incarceration, such as spending time with old friends, searching for easy money, participating in side-relations and one-night stands, were more prone to re-offend than those who socially isolated themselves and participated in more pro-social behaviour or intimacy behaviours (Seal, Eldrige, Kacanek, Binson & Macgowan, 2007: 2398. Nevertheless, for some criminals, when released from prison, the possibility of entering criminal networks can seem to be one of their favourite options, even if they realise that this can be self-defeating (Scott, 2004: 74).
1.5.1.7 Prior criminal history Prior illegitimate record, including the aggregate of preceding arrests, prosecutions and the duration of the first crime or sentence, has not only been reliably related to recidivism in empirical studies but has also proven to be a strong correlation to recidivism. Pritchard (1979:27) analysed 71 recidivism trials, including 177 separate surveys of criminals, and found that the prevalence and amount of previous adult convictions contributed to recidivism in 99 of the 116 cases in which it was studied, while the age at first indictment was linked to recidivism in 77 of the 95 studies which examined its effect on recidivism. In contrast, Burgoyne (1979:96) found that the number of former convictions and age at first arrest was the greatest predictor for recidivism in a study of criminals discharged from Victorian jails between January 1972 and December 1973. Furthermore, a higher rate of recidivism was observed for those criminals with a higher number of prior convictions and those whose first arrest came at an early age.
Similar results were recorded if the previous record of the crime was specified in terms of the number of past adult arrests (Bureau of Justice Statistics, 1989: 13). The study found that the more severe a pre-arrest inmate released, the lower his or her possible rate of recidivism. Of those released inmates with only one previous adult conviction, 38.1 percent were re-arrested during the three-year follow-up cycle relative with 82.2 percent of those released inmates with 16 or more prior adult convictions. The percentage of recent adult convictions was a strong predictor for recidivism even when the age of release from prison, sex and race and the number of previous incarcerations (Illinois Criminal Justice Information Authority, 1985: 29) was taken into account.
Empirical studies have found that prisoners who have a past term of imprisonment (United States Bureau of Justice Statistics 1984Statistics :22, 1989Burgoyne 1979:39), have been state-owned or have received a number of concurrent parole orders (Burgoyne 1979:39) have a high level of recidivism, indicating that existing punishment and criminal record are also a source of recidivism. Throughout his dissertation on "Seeing Class Self," Cooley (1998: 490) explains that the idea of self-concept is a function of other perceptions towards the person concerned. If others deal with an individual as if he were particular characteristics, then a self-fulfilling prophecy is created. So if other people think that we are such an individual (smart, educated, nice, respectful and criminal), then they act accordingly. Our self-concept and behaviour are therefore formed by a tag. The use of the tag on offenders as "criminals" or "evil" people was meant to prevent violence but, sadly, the unanticipated effect of the mark is the continuation of the offense. Therefore, the more negatively branded an individual, the greater the predisposition to commit more crime. (Akers, 1997: 40).
It is important to note that these marks reflect what Braithwaite (1989: 159) considered disintegrative guilt because they were directed at stigmatizing or condemning the individual as an undesirable member of society. The perception of detention prohibits inmates from having multiple resources and opportunities to avoid reoffending (Malott & Fromader, 2010: 521). Background checks are the gate-keeping tool used by companies to weed out candidates with or without criminal records. According to Petersilia (2003: 87), employment opportunities legally restricted for ex-offenders in America include childcare, schooling, safety, nursing, and home health care. She states that, in a State like California, ex-offenders are legally prohibited from certain occupations such as business, real estate, pharmacy, counseling, physical therapy, and health.
In a specified follow-up period, a criminal with prior contact with law enforcement officers does not become completely indistinguishable from those without prior contact with respect to the risk of offending (Kurlychek et al., 2006: 309). The more a person lives a crime-free life, the more he or she understands the value of criminal isolation. In contrast, criminals with strong criminogenic conditions re-offend more than those who try to avoid fresh crimes. Notwithstanding that, it is an open secret that it is difficult for people with a criminal history to secure employment.

Recidivism in Kenya
Article 28 of the Constitution of Kenya provides that every person has inherent dignity and the right to have that dignity respected and protected (Republic of Kenya, 2010: 25). In addition, Article 51 of the Kenyan Constitution provides that a person who is arrested, held in custody or imprisoned under the law shall maintain all the privileges and basic freedoms set out in the Bill of Rights, except to the degree that any specific right or fundamental liberty is explicitly inconsistent with the condition that the person is detained, held in custody or imprisoned. (Republic of Kenya, 2010: 36). The law further provides for the humane treatment of prisoners in line with the United Nations Standard Minimum Rules for the Treatment of Prisoners to which Kenya is a signatory. These legal provisions are entrenched in the Prisons Act (Chapter 90) laws of Kenya which emphasizes on humane treatment of offenders.
Released prisoners in Kenya have a seventy-five percent risk of committing another crime and a fifty percent probability of going to jail two years following their release from prison (Oruta, Omosa & Lumumba, 2017:101). This phenomenon compounds the high prison population problem and overcrowding. The extremely high rate of recidivism has immense costs in terms of public safety and money spent on investigating, punishing and incarcerating re-offenders. The next sections highlight the various correctional institutions and programmes in Kenya that take part in offender management including recidivists.

Theoretical underpinnings of recidivism
Recidivism is not explained by a single theory (Ryan & Yang, 2005: 186). Correlations in relapsed offenders' recidivism and criminal activity are the product of a complex phenomenon. Strain theories and labeling theory are used in the study. One model, positive psychology is also used. These theories and model explain the relationship between socio-economic, human and environmental factors and recidivism.

Strain theories
In this category of strain theories, two theories are discussed, namely: the General Strain Theory by Robert Agnew and the Strain Theory by Robert K. Merton. In 1992, Robert Agnew developed the General Strain Theory, partly as a response to the disadvantages and limitations of older strain models that were almost abandoned in the latter part of the 20th century (Agnew, Brezina, Wright, & Cullen, 2002: 334;Slocum, Simpson, & Smith, 2005: 468). Nevertheless, Agnew's strain model has origins that could probably trace back to the turn of the 19th century, when Emile Durkheim published his notorious novel, "Suicide" (Durkheim, 1897: 597). Durkheim's emphasis on self-destruction and suicide led him to the idea of anomie which he defined as a state of normality that could lead to a lack of norms or guidelines for people's behaviour (Durkheim, 1897: 693). As with Agnew's general theory of strain, Durkheim's concept of anomie is also quite direct, especially when looking at his explanation of anomie. Throughout his study of suicide, Durkheim speaks also about the influence of crises, and how crises reflect dislocations and irregularity throughout one's existing existence, forcing people into unusual or unknown circumstances (Durkheim, 1897: 699). Being in this state without rules or expectations as to what is appropriate or feasible can cause people to lose understanding and lead them to a state of dissatisfaction and torment triggered by their pursuit of unattainable goals and capabilities (Akers & Sellers, 2004: 317).

1.5.3.2The Labeling theory
Labeling theory is founded on the premise that some members of society are capable of building and applying attributes to other members of the same society (Becker, 1963: 207). According to research, the application of a negative label by one social group to another produces another and thereby stigmatises the individual or group to which the tag has been applied and is considered beyond traditional society (Akers & Sellers, 2009: 211;Becker, 1963: 201). Becker (1963: 201), the founder of labeling theory, stated that deviant behaviour only exists after members of society have defined it as such.
Labeling theory falls within the symbolic interactionist paradigm which assumes that one's identity and selfconcept are continuously determined by interactions with others and thus exist only on the basis of social interaction (Akers & Sellers, 2009:169). This can, therefore, be concluded that those individuals who are negatively branded would incorporate this tag into their view of themselves. Goffman (1963: 73) assumed that those who were branded would not act in ways that undermined the tag, but rather display actions that validated it. Akers and Sellers (2009: 188) say a person will face humiliation and shame once it has been branded. It is these feelings that will provide motivation to engage in further deviant acts for labeled individuals. Furthermore, once labeled, the label recipients adopt the characteristics generated as part of their primary identity and live in ways that confirm the stereotypes attached to the label, thereby confirming their authenticity to the individual.

Study Findings
Study findings are presented in this section;

Relationship between offender characteristics and recidivism
The first specific objective of the study seeks to investigate the relationship between offender characteristics and recidivism among released prisoners in Kakamega County.  .000 N 329 Source: Field data, (2018) ** Correlation is significant at the 0.01 level (2-tailed). Study findings in table 2 reveal a significant relationship between individual characteristics and recidivism among released prisoners in Kakamega County (r=0.669; P< 0.01). This implies that individual characteristics of respondents have a significant influence on repeat offending among released inmates in Kakamega County. The null hypothesis that states that there is no significant relationship between individual characteristics and recidivism among released prisoners is therefore rejected at the level of significance of 0.01 and its alternative which states that there is a significant relationship between individual characteristics and recidivism among released prisoners adopted.
To determine the differences in the extent of the influence of individual characteristics on recidivism, measures of dispersion and variability are computed and findings presented in table 3. Study findings in table 3 reveal that the mean for gender is the highest, namely 1.39. This implies that gender is the single individual characteristic with the highest influence on recidivism. There is a significantly higher number of male recidivists compared to incarcerated male offenders as compared to female recidivists as compared to incarcerated female offenders.
The age of respondents has the second-highest influence on recidivism with a mean of 1.37. This also reflects the age differences among recidivists, since there are more youthful offenders in prison as compared to aged or older offenders.
Offence type has a mean of 1.31 which points to the influence of offence type to recidivism. Offences against property are more prevalent among sampled recidivists as compared to offences against persons. Drug and substance abuse has a mean of 1.27 implying that even though there are recidivists who have committed offences related to drug and substance abuse, the rate of recidivism in this category of offence is not prevalent. Prison sentence as long, medium or short has a mean of 1.18 implying that the length of a prison sentence has significantly minimal influence on recidivism. Given the small differences in the means for the various individual characteristics in explaining recidivism, there is a need to establish whether these differences in the means are statistically significant. In this regard, a onesample independent t-test for equality of means has been computed at 0.05 level of significance and findings presented in table 4. Study findings in table 4 indicate that there is a statistically significant difference in the mean between the various individual characteristics as indicators of recidivism among released prisoners in Kakamega County (t=3.308, P < 0.05, df= 2, 28). This is further shown where the critical value of t (2.99) is less than the calculated value of t (3.308).

Regression analysis of offender characteristics and recidivism
Research data on offender characteristics has been subjected to regression analysis to predict recidivism amongst offenders released from prisons within Kakamega County and findings presented in table 5. Study findings from regression analysis where offender characteristics are used as predictors of recidivism reveal an R squared value of 0.237 implying that offender characteristics account for 23.7% of the variance in recidivism among sampled offenders.

Interviews with Probation and Prison Officers
Interview guides were used to seek the opinion of Probation Officers and Prison Officers on the relationship between offender characteristics and recidivism. The aim is to assess whether the characteristics exhibited by offenders would explain the possibility of repeat offending. The majority of the interviewed Probation Officers (69.4%) were of the view that offender characteristics have a significant influence on recidivism among released prisoners in Kakamega County.
The majority of the interviewed Probation Officers (57.1%) were of the general view that more youthful offenders were more likely to breach Probation Orders as compared to older offenders. Probation Officers were also of the view that male offenders were highly likely to breach the conditions of the Probation Orders as compared to their female counterparts. In all cases where an offender breaches Probation Orders, they get arrested for the breach of the Order and an alternative sentence is meted out for them hence making them repeat offenders. In Kenya, in the year 2017, there were 16 371 male recidivists out of an estimated 52 000 male Most of the interviewed Probation Officers (52.9%) also indicated that the type of offence committed had a significant influence on recidivism. Officers who were interviewed noted that offenders who had committed crimes that were utilitarian in nature such as theft, burglary, etc were more likely to repeat similar offences upon release from prisons or more severe offences such as attempted robbery or robbery. This corroborates with the Kenya Economic Survey (2018:270), which reports that the following crimes were committed by convicted offenders in 2017  Order and administration of lawful authority 8 505 The majority of interviewed Prison Officers (72.3%) were of the view that offenders sentenced for petty offences were highly likely to recidivate due to the nature of the short sentences that they received, with most of them ranging from one week to six months. During this period, offenders are not likely to undergo any tangible rehabilitation programme. Further still, offenders with drug-related offences or those with a history of narcotic drug use were more likely to engage in repeat offending given the negative influence of the drug and also given the peer group associations that come with drug use. From the above response from correctional officers, it is evident that individual characteristics of an offender have a significant role in determining recidivism among released prisoners.

Focus group interviews
The general thread emanating from the focus group discussions seems to allude to mixed findings for different offenders. Some family and community members (49.8%) were of the view that the offenders have improved in character after incarceration while others (50.2%) were of the view that offenders have worsened in character. Other community members (2.7%) opined that there was no significant change in the character of the offenders before and after incarceration or placement on probation.
Focus group discussions with two ex-offenders, four family members of an ex-offender and a victim of crime, four community members as well as local administration, and a religious leader presented mixed feelings regarding the potential of offenders to engage in repeat offending. Even though family members of ex-offenders were optimistic that ex-offenders would eventually change and become law-abiding, community members and local administrators held divergent views regarding the possibility of ex-offenders reforming after a period of incarceration.
The above discussions allude to different viewpoints for different categories of community members of recidivists. In as much as close relatives to recidivists might want to portray offenders as being capable of reforming and become good citizens, community members seem to paint offenders as people who cannot change and who should be suspected of any wrongdoing in the community when there is nobody else to suspect.

Recommendations
This section presents the recommendations of the study based on the findings of the investigation. Offender characteristics and recidivism The study reveals a significant relationship between offender characteristics and recidivism. From the study, it is recommended that treatment plans drawn by correctional officers should take into account the individual characteristics of released offenders since there is a strong association between individual characteristics of offenders and recidivism. The government should facilitate the correctional officers with the required resources to use the Actuarial Risk Assessment model. The model is applied to released offenders to predict the future probability of recidivism (Robinson & Crow, 2009: 91). This can be achieved through the application of the "Offenders Group Reconviction Scale" a windows based programme for use by correctional officers. This involves a database consisting of information about the demographic characteristics and offending histories of the offenders. The instrument provides an estimate expressed as a percentage of the statistical likelihood of reconviction within two years of release from custody (Robinson & Crow, 2009: 91). The key variables that Offender Group Reconviction Scale considers in calculating the statistical likelihood of reconviction are  Age;  Gender;  Offense type;  Numbers of previous convictions; and  Age at first conviction. These are key constructs identified by the study to have a significant correlation on recidivism. The results obtained from the instrument can be used to come up with the necessary categorisation of offenders based on their risk levels of re-offending. Correctional officers' ability to classify offenders into "low-risk" and "highrisk" groups will enable them to develop offender treatment plans that are effective based on the risk levels of individual ex-offenders.

Gender
The study established that there is a significantly higher number of male recidivists compared to the total male prison population than that of female recidivists computed against the total female prisoners. From the study, it is recommended further investigations to establish the main causes of gender differences in re-offending. The study also recommends a gender-sensitive treatment approach to the rehabilitation of offenders. This is because the current generalised approach to the treatment of offenders has been pointed out to be ineffective in addressing rehabilitation needs among male recidivists.

Age at first conviction
The study reveals a significant relationship between age at first conviction and recidivism. Because incarceration is particularly prevalent among youthful offenders as found by the study, the impact on their future adult life will be more pronounced since they are wasting the energetic part of their life in prison. It is recommended that that the government and other stakeholders come up with policies and programmes aimed at specifically addressing factors influencing increased youth reoffending. In addition, extensive post-release treatment and after-care services for youthful offenders need to be developed, implemented, monitored and regular assessments are done to review the progress made in reducing recidivism.

Offence type
Offences against property and drug-related were more prevalent among recidivists. It is recommended that the offenders should take individual responsibility and make a conscious decision to desist from crime and drug abuse. In addition, there is the need for correctional officers to introduce evidence-based treatment of offenders with a specific focus on those convicted against property and drug-related offences as a way to mitigate against the prevalence of the two forms of crime amongst released offenders.

Drugs and substance abuse
The need for alcohol and drug treatment should be addressed amongst released offenders. Drug and alcohol abuse is implicated in the crimes and incarceration of the majority of prisoners. There's a strong link between possession and sales of drugs and offences committed by ex-offenders to obtain money to purchase drugs e.g. burglary and robbery. Prison-based and community-based drug treatment programmes should be emphasised in sustaining sobriety. Ex-offenders should be encouraged to form self-help groups to help them with long-term social support that aids them in successful reintegration into the community.

Length of the prison sentence
Since the investigation did not find a significant influence of the length of a prison sentence on recidivism, it is recommended that prison-based rehabilitation should take into account the imprisonment duration of offenders.