Violence in youth custody: Risk factors of violent misconduct among inmates of German young offender institutions

Abstract

Inmate-on-inmate violence is a serious problem in prisons and young offender institutions. However, most studies on predictors of violent misconduct have focussed on adult inmates. This study examines the perpetration of violence in multiple young offender institutions, using the self-report data of 865 male inmates. Prevalence rates indicate that violence occurs to a high extent in the institutions. Regression analyses show that both importation and deprivation variables significantly predict the perpetration of physical and sexual violence. Specifically, drug use during imprisonment, violent beliefs, and a negative inmate–staff relationship were found to increase the risk of violent misconduct. The implications of these findings are discussed.

Keywords

Young offenders prison violence risk factors importation model deprivation model

Introduction

Violent misconduct in prison, including physical and sexual assaults against other inmates, is a common problem in both adult prisons and young offender institutions (YOIs). Prevalence rates differ considerably, depending on the methodology, the sample, and the definition of prison violence employed in the research (see, e.g., Ireland, 1999; Tasca et al., 2010). Irrespective of the exact frequency of inmate-on-inmate violence, its occurrence is a major concern. Prison violence not only needs to be prevented because inmates have a right to be treated humanely, even while imprisoned, which includes protection from victimization. Prevention of violence, especially among young inmates, is also important because it can impact future offending after release (Trulson et al., 2005), which constitutes a threat to public safety.

What has to be noted is that most research on inmate-on-inmate violence focusses on samples of adult offenders. It is questionable if the findings from these studies can be generalized to juvenile and adolescent inmates, because their developmental trajectories are still evolving (cf. Moffitt, 1993; Sampson and Laub, 2003). Additionally, the vast majority of studies on the topic of inmate-on-inmate violence have been conducted in the US and in the UK. It is not clear if results from research in these countries also apply to other prison populations in different countries with different legislations and conditions of confinement.

Thus, the aim of the present study is to close this research gap by examining predictors of prison violence among young offenders in Germany. Self-report data were obtained from 865 male inmates to assess which personal and institutional factors are associated with the perpetration of physical and sexual inmate-on-inmate violence.

Theoretical and empirical background

Previous research indicates that violence among juvenile and adolescent inmates occurs rather frequently. Power et al. (1997), for example, reported that almost 30 percent of the 707 juvenile inmates they surveyed had bullied other inmates since they were in the institution. A study of violent victimization among inmates of a German YOI found that 34.9 percent of the inmates had been physically assaulted while imprisoned and 53.3 percent of the young offenders indicated that they had been physically threatened during their current sentence (Kury and Smartt, 2002). Tasca et al. (2010) conducted in-depth interviews with 95 Black and Latino offenders in the US, aged 16–18. More than half of the participants indicated that they had assaulted another inmate at least once during the 12 months preceding the interview. Using a self-report behavioural checklist, Ireland (1999) found that among a sample of adult and young offenders of both sexes in the UK, 37.5 percent stated that they had used direct bullying (e.g. physical or sexual bullying) at least once against another inmate.

Different theoretical approaches have been brought forward to explain violent misconduct among prisoners. The traditional and possibly most intensively studied ones are the importation and deprivation models. The importation model suggests that inmate (mis-)conduct reflects the attitudes, beliefs, and behavioural patterns that the inmate has acquired prior to his or her incarceration (Irwin and Cressey, 1962). The deprivation model, on the other hand, suggests that inmate behaviour can be best explained by the conditions of imprisonment itself, that is, the “strains” of life in prison to which the inmate must adapt (Sykes, 1958). Another current model of prisoner misconduct is the management theory (DiIulio, 1987), sometimes also referred to as the administrative control theory (e.g. Hochstetler and DeLisi, 2005; Huebner, 2003). The management theory posits that individual as well as collective inmate misconduct (i.e. prison riots) is a consequence of poor prison management. Poor prison management is characterized by factors such as security lapses and a lack of discipline amongst staff.

Several studies have examined the association of various importation, deprivation, and management factors with prison violence. Overall, the findings indicate that all three theories provide a useful basis for explaining violent behaviour among inmates and the best prediction of prison violence can probably be achieved by integrating multiple models (cf. Harer and Steffensmeier, 1996; Hochstetler and DeLisi, 2005; Huebner, 2003; Jiang and Fisher-Giorlando, 2002; Lahm, 2009; MacDonald, 1999).

Importation factors

Research has uncovered a number of variables that are associated with violent behaviour among inmates. Among the importation factors that have been found to be related to prison violence are age, ethnicity, criminal history/prior offenses, drug use, and criminal or violent beliefs. Age has consistently been found to be negatively correlated with violent misconduct, indicating that younger inmates are particularly likely to engage in physical assaults against other inmates. Kuanliang et al. (2008), for example, found that among a large sample of male inmates in the state prison system, juveniles exhibited higher rates of violent misconduct and assault than adult offenders. Furthermore, the difference between young and adult offenders increased with the increasing severity of misconduct. Inmates aged 15 exhibited the highest rate of assaults. The inverse relationship of age and prison violence was also found in other studies using samples of both male and female inmates (Cunningham and Sorensen, 2007; Griffin and Hepburn, 2006; Sorensen and Cunningham, 2007, 2010).

Further studies have found that members of ethnic minorities are more likely to engage in violent misconduct in prison than white inmates (e.g. DeLisi et al., 2004; Harer and Steffensmeier, 1996). Furthermore, Berg and DeLisi (2006) reported differences in the level of prison violence between several ethnic groups. They found that Hispanic and Native American males exhibited the highest rate of violent misconduct, whereas Native Americans and African Americans were the most violent among the female inmate sample. However, other researchers did not find ethnicity to be significantly related to inmate-on-inmate violence when controlling for the effects of other variables (Jiang and Fisher-Giorlando, 2002; MacDonald, 1999; Sorensen and Cunningham, 2007; Tasca et al., 2010). Citizenship was also not found to be a significant predictor of violent behaviour against fellow inmates (Berg and DeLisi, 2006).

Research focussing on prior offences has produced strong evidence that inmates with a history of violent offences have an elevated risk of engaging in prison violence (e.g. Berg and DeLisi, 2006; DeLisi et al., 2004; Griffin and Hepburn, 2006; Harer and Steffensmeier, 1996; MacDonald, 1999).

Findings on the influence of substance abuse are inconsistent; some studies indicate a positive association of drug use and alcohol dependence with prison misconduct (Flanagan, 1983; Jiang, 2005; Mills et al., 1998), whereas MacDonald (1999) reported that juveniles who have committed drug-related offences have a reduced risk of engaging in inmate-on-inmate violence. These inconsistent results might be due to different methodologies of the studies.

Further research has focussed on violent or criminal beliefs and found that such beliefs are significantly associated with prison violence. Hochstetler and DeLisi (2005), for example, reported that criminal attitudes were positively related to prison offending among a sample of 208 male work-release parolees.

Deprivation factors

Different measures of deprivation have been studied as a potential influence on the occurrence of prison violence. Factors such as prison population size and crowding have been found to be positively related to inmate-on-inmate violence in some studies (Gaes and McGuir, 1985; Lahm, 2008). Franklin et al. (2006) conducted a meta-analysis on the effect of overcrowding on inmate misconduct and found that, overall, prison crowding has no significant impact on misconduct.

A further variable found to influence violent misconduct among inmates is the relationship of staff and inmates. Harer and Steffensmeier (1996) reported that a high proportion of staff who have positive attitudes towards inmates is negatively associated with prison violence. This result might be related to the finding that prisoners experience less strain if prison staff have a favourable orientation towards working with the inmates (Molleman and Leeuw, 2012). Liebling (2011) and Crewe et al. (2011) have also stressed the importance of a good relationship between prison officers and inmates in maintaining order and thus preventing prison violence. Research by Liebling (2008) has shown that inmates who think that they are treated unfairly by the prison staff tend to become insubordinate, which might lead to violent misconduct.

Regarding the influence of prison type on offender misconduct, research has produced inconsistent results. Davidson-Arad (2005) compared three types of juvenile correctional facilities in Israel and found that violent misconduct was more frequent in hostels than in diagnostic and closed facilities. Another study has found inmate-on-inmate violence to be positively related to the security level of the institution (Harer and Steffensmeier, 1996), but there is also research that failed to find any effect of prison type on inmate misconduct (Camp and Gaes, 2005).

In summary, several deprivation variables have been uncovered that have a potential influence on levels of prison violence. These variables are likely to differ between prison facilities. Therefore, differences in the rates of violent misconduct between prisons are likely to occur.

Management factors

Indicators of poor prison management have been found to be associated with violence and misconduct in prisons. Previous research has shown that prisoners’ experience of coercive control was positively related to violent misconduct (Day et al., 2015). Reisig (1998) also found that an overly strict and controlling management style was associated with elevated levels of prison disorder, compared to management practices that focus less on formal control mechanisms. Sparks and Bottoms (1995), as well as Liebling (2008, 2011), have stressed the importance of fair procedures and legitimate exercise of control and authority in prison to avoid inmate misconduct in general and violence in particular. Perceived legitimacy and fairness are strongly influenced by inmate–staff interactions and thus, linked to the deprivation variable of inmate–staff relationship.

Aim of the present study

Research has uncovered a number of importation, deprivation, and management factors that influence prison violence, although findings regarding some of the variables are inconsistent, which might be due to different samples and methodologies used by the researchers. To date, only a few studies have investigated the extent of inmate-on-inmate violence in Germany (e.g. Ernst, 2008; Wirth, 2007). Wirth (2007) analysed official records of violent incidents in prisons in North Rhine-Westphalia. The sample consisted of 638 inmates. Only descriptive statistics are reported and the concurrent effects of risk factors were not considered in the study. Ernst (2008) obtained questionnaire data from male inmates in 33 German prisons. The results are focussed mostly on the prevalence of inmate-on-inmate violence and, again, only descriptive statistics are provided instead of multivariate analyses.

To the authors’ best knowledge, only two studies have explicitly examined prison violence in German YOIs (Kury and Smartt, 2002). The study conducted by Kury and Smartt (2002; Neubacher, 2014). was based on questionnaire data from 168 male inmates. The results are descriptive in nature and focussed predominantly on pre-prison victimization as a correlate of inmate-on-inmate violence. The simultaneous influence of multiple risk factors was not examined. A longitudinal study by researchers from the University of Cologne examined among other things the occurrence of suicide and suicidal thoughts as well as inmate-on-inmate violence in YOIs in Thuringia and North Rhine-Westphalia, focussing on physical and psychological violence, damaging of goods, and coercion (Neubacher, 2014; Wolter and Boxberg, 2016).

The aim of the present study is to collect data on the prevalence of the perpetration of physical and sexual prison violence in multiple German YOIs. Furthermore, the study seeks to examine which importation and deprivation factors significantly impact inmate-on-inmate violence among the young offenders, simultaneously considering the influence of multiple risk factors. Information on which risk factors predict prison violence in German YOIs can then be utilized to develop effective means to reduce the occurrence violence among inmates.

We were not able to directly analyse the effects of management factors on prison violence, because our analyses are solely based on inmates’ self-report data, which did not include questions regarding the YOIs’ managerial practices. However, a number of variables on which data were collected are related to the prisons’ management style and this association will be highlighted in the respective paragraphs.

Method

Sample and data

In 2011 and 2012, the Criminological Research Institute of Lower Saxony conducted a quantitative survey among inmates of five German YOIs. In Germany, the term “young offenders” refers to people who are sentenced under the Juvenile Court Act, which applies to people who are at least 14 years old at the time they committed the crime. Offenders can be sentenced under the Juvenile Court Act up to an age of 21. The age of young offenders imprisoned in YOIs ranges from 14 to 25 years. Self-report data were used, because they have been found to be more reliable than official records, probably due to a reluctance of inmates and prison staff to report misconduct (cf. Hewitt et al., 1984).

The survey was conducted in four of the 16 federal states (Bundesländer) of Germany, namely Brandenburg, Lower Saxony, Saxony, and Thuringia. The survey started in Lower Saxony, where the Criminological Research Institute is located. The study was then rolled out to the nearby federal states of Brandenburg, Saxony, and Thuringia. In Lower Saxony, questionnaires were administered in two YOIs, one being a large correctional facility that can accommodate up to 645 young offenders. The second YOI is a smaller prison with room for a maximum of 97 prisoners. All male young offenders in Lower Saxony who are imprisoned for the first time and serve a sentence of three and a half years or less are detained in the latter YOI. All other young offenders in Lower Saxony are imprisoned in the larger correctional facility. The YOI in Brandenburg has room for up to 198 male prisoners who serve a sentence of between six months and 10 years. Saxony’s YOI has a maximum capacity of 347 and accommodates all male young offenders in Saxony. The YOI in Thuringia can accommodate 183 male young offenders.¹

Overall, 865 male young inmates completed the questionnaire, which corresponds to an average return rate of 71.9 percent. An exclusively male sample was used for the study, because prevalence rates of inmate misconduct are much higher for males compared to females (Ireland, 1999; Wirth, 2007).

Participants completed the questionnaire alone and were at no point asked to provide their name or any other personal information that could be used to identify them. Completed questionnaires were placed in a sealed envelope by the participants and the envelopes were not opened before they had been sent collectively to the Criminological Research Institute of Lower Saxony.

The respondents’ age at the time they were surveyed ranged from 14 to 25 years. The majority of participants (70.2%) were younger than 22 years. About two thirds of the sample were German, whereas 32.1 percent of the inmates had a migration background. Participants who were not of German origin came predominantly from Russia, Turkey, or the Former Republic of Yugoslavia.

A total of 69.4 percent of the respondents were serving a prison sentence for violent or sexual crimes. Participants convicted for property crimes (e.g. fraud or theft) constituted 22.7 percent of the sample.

Variables and measurement

Several importation and deprivation variables that have been found to be predictive of inmate-on-inmate violence were measured in the survey.

Dependent variables

Participants were asked to indicate on a 4-point scale how often they had perpetrated different forms of physical and sexual inmate-on-inmate-violence during the last four weeks (for a list of the items, see Appendix 1). The items used are a subset of items from the Direct and Indirect Prisoner Behaviour Checklist-Scaled Version Revised (DIPC-SCALED-r ©Ireland, 2007), supplemented with items developed by researchers of the Criminological Research Institute of Lower Saxony. We did not include the complete DIPC-SCALED-r to keep the questionnaire relatively short. This was done to reduce the risk of participants dropping out of the study. The items were translated into German by researchers of the Criminological Research Institute (see Baier and Bergmann, 2013). We wanted to focus on direct physical and sexual prison violence and thus, excluded items concerning psychological/verbal forms of bullying, such as calling someone names. The scale ranged from “never” to “often”. A reference period of four weeks was chosen instead of a longer timeframe to reduce the risk of memory distortion. Since the majority of participants indicated no perpetration of prison violence or disclosed engaging in violence only rarely, the perpetration of physical and sexual violence was coded dichotomously (no perpetration vs. at least one incident of prison violence during the last four weeks).

Independent variables

Participants were also asked to indicate their age by selecting one of three age categories. Additionally, the young offenders were asked to state their citizenship and the native country of their parents. Inmates were considered as having a migration background if they had a citizenship other than German or if their parents were from a country other than Germany. Because the numbers of inmates of any particular ethnic origin except German were very small (all n’s ⩽ 50), we used a dichotomous variable for the statistical analyses (German origin vs. migration background).

The questionnaire did not include questions regarding the participants’ pre-prison substance abuse. Therefore, we used drug use during imprisonment as a proxy for the importation variable of pre-prison drug use, because previous research has shown that drug use in prison is usually a continuation of pre-prison substance abuse (Keene, 1997; Thomas, 1977; Thomas and Cage, 1977). Consequently, drug use is most likely a factor that the inmate “imports” into the prison system. The variable was coded dichotomously, that is, no use of drugs vs. at least one incident of drug use during the last four weeks. Again, a timeframe of four weeks was chosen instead of a longer period to acquire reasonably accurate data. To measure drug abuse during imprisonment, we used three items from the DIPC-SCALED-r. Participants were asked if they had smoked cannabis, injected drugs, or used drugs other than cannabis.

Regarding the criminal history of the inmates, we asked participants to indicate whether they were serving their current prison sentence for a property offence (e.g. burglary), a violent offence (e.g. assault), a narcotics offence, a sexual offence, or an offence of a different type. For the analyses of prison violence, we differentiated between inmates serving a sentence for a violent offence (i.e. a violent or a sexual offence) vs. a non-violent offence (i.e. all other offences).

The final importation factor considered in this study refers to the inmates’ attitude regarding violence, that is, violence approval. This variable was measured using six items from the Measures of Criminal Attitudes and Associates (Mills et al., 2002). Each item was measured on a scale ranging from 1 to 4 with greater values indicating a stronger approval of violence (Cronbach’s α = 0.91). Items used are, for example, “Someone who makes you very angry deserves to be hit” and “It is not wrong to fight to save face”.

The deprivation variable “Relationship between prison staff and inmates” was assessed using six items from the Measuring the Quality of Prison Live Survey (Liebling assisted by Arnold, 2004; see also Crewe et al., 2015). The 6-item scale used in our study ranged from 1 to 4 (Cronbach’s α = 0.86), with greater values indicating a more positive inmate–staff relation. The scale included items such as “I trust the officers in this prison” and “Relationships between staff and prisoners in this prison are good”. Furthermore, we dummy-coded the facility where the participants served their prison sentence as a proxy for the deprivation model, because the prison environment, security measures, daily routines etc. are likely to differ between the different YOIs.

The two deprivation variables are also linked to the management model, because managerial practices probably influence the prison officers’ behaviour. Additionally, it can be assumed that the prison management style differs between the five facilities.

In the following, descriptive statistics are presented as well as correlations between the variables. Finally, we present binary logistic regression models to examine which importation and deprivation variables significantly predict physical and sexual inmate-on-inmate violence.

Results

Descriptive statistics and bivariate analyses

About one third of the participants (33.0%) indicated that they have physically victimized fellow inmates during the four weeks preceding the survey. A comparably low proportion of the inmates surveyed (3.6%) indicated that they have been the perpetrator of sexual violence during the last four weeks (see Table 1). Descriptive statistics of the importation and deprivation variables used to predict the perpetration of physical and sexual prison violence are also presented in Table 1. Almost one third (30.2%) of the inmates indicated using drugs at least once during the four weeks preceding the survey. Participants’ average score on the “violence approval” scale was 2.32 (SD = 0.92); the mean score on the “Relationship with staff” scale was 2.59 (SD = 0.80). Further descriptive statistics have already been presented in the “Sample and data” section.

Table 1.

Descriptive statistics (N_max = 865 young offenders).

Variable	%	M (SD)
Dependent variables
Physical violence	33.0
Sexual violence	3.6
Independent variables
Facility
● Facility 1	43.4
● Facility 2	6.1
● Facility 3	14.9
● Facility 4	27.4
● Facility 5	8.2
Age
● 14–17 years	6.9
● 18–21 years	63.3
● 22–25 years	29.8
Migration background	32.1
Criminal history (violent offence)	69.4
Drug use during the last 4 weeks	30.2
Violence approval		2.32 (0.92)
Inmate–staff relationship		2.59 (0.80)

Note: Violence approval and inmate–staff relationship were measured on a 4-point scale, with greater values indicating stronger violent beliefs and a better inmate–staff relationship, respectively.

To assess if prison violence is related to the importation and deprivation factors examined in this study, Spearman’s rho was calculated. We used non-parametric correlations, because some of the variables were not normally distributed or dichotomous. Correlation coefficients are presented in Table 2. A number of significant associations were observed. Physical violence correlates most strongly with drug use (r_s = .41, p < .001) and violence approval (r_s = .38, p < .001); sexual violence is most strongly related to drug use (r_s = .23, p < .001) too. The two measures of prison violence are also positively and significantly correlated with each other (r_s = .23, p < .001). The importation and deprivation variables used to predict inmate-on-inmate violence are also in part correlated with each other, but the correlation coefficients are generally small (all r_s’s < .30).

Table 2.

Spearman’s rho correlations between the importation and deprivation variables as well as the outcome variables of physical and sexual prison violence.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
(1) Age
(2) Migration background	−.12^b
(3) Criminal history (violent offence)	−.08^a	.01
(4) Drug use	−.03	.12^b	.13^c
(5) Violence approval	−.09^a	.12^b	.20^c	.29^c
(6) Inmate–staff relationship	.09^a	−.07	−.05	−.24^c	−.21^c
(7) Physical violence	−.04	.10^b	.13^c	.41^c	.38^c	−.23^c
(8) Sexual violence	.00	−.00	.01	.23^c	.13^c	−.13^c	.23^c

p < .05.

p < .01.

p < .001.

We also examined if violent misconduct among young inmates varied between the five YOIs. As can be seen from Table 3, prevalence rates of physical perpetration differ markedly between prisons, ranging from 15.9 percent (Facility 5) to 38.8 percent (Facility 1). These differences in physical perpetration between the facilities reached statistical significance (χ²(4) = 22.29, p < .001). Rates of sexual violence ranged from 1.4 percent (Facility 5) to 6.5 percent (Facility 4).

Table 3.

Prevalence rates of physical and sexual violence (perpetrators) for five German YOIs.

	Facility 1	Facility 2	Facility 3	Facility 4	Facility 5	Total
Physical violence	38.8%	24.5%	23.8%	36.7%	15.9%	33.0%
Sexual violence	3.2%	1.9%	1.6%	6.5%	1.4%	3.6%

Multivariate analyses

To assess which importation and deprivation factors significantly influence the perpetration of physical and sexual prison violence among inmates of German YOIs, while controlling for the influence of other factors, binary logistic regression analyses were performed (see Table 4).

Table 4.

Binary logistic regression analyses of importation and deprivation factors predicting perpetration of physical (Model 1) and sexual prison violence (Model 2).

Predictor	Model 1	Model 2
	(Odds ratio)	(Odds ratio)
Facility 1	Reference	Reference
Facility 2	0.69	1.77
	(0.43)	(1.13)
Facility 3	0.83	1.01
	(0.29)	(0.85)
Facility 4	1.40	3.54^a
	(0.24)	(0.53)
Facility 5	0.63	2.01
	(0.42)	(1.14)
Aged 14–17 years	Reference	Reference
Aged 18–21 years	1.05	0.44
	(0.40)	(0.73)
Aged 22–25 years	1.08	0.75
	(0.42)	(0.78)
Migration background	1.11	0.87
	(0.21)	(0.51)
Criminal history (violent offence)	1.28	0.59
	(0.21)	(0.51)
Drug use	4.24^c (0.20)	10.21^c (0.59)
Violence approval	2.04^c	1.73
	(0.11)	(0.28)
Inmate–staff relationship	0.68^b	0.55
	(0.13)	(0.31)
Constant	0.11^c	0.01^b
	(0.62)	(1.52)
Cox–Snell R²	0.24	0.07
Nagelkerke R²	0.33	0.27

Note: Due to missing data, the sample size was reduced to n = 709. Standard errors are presented in parentheses.

p < .05.

p < .01.

p < .001.

The independent variables included in the analyses are age (14–17 years (= reference category), 18–21 years, and 22–25 years), migration background (German origin = 0, other ethnic origin = 1), criminal history (non-violent = 0, violent = 1), drug use during the last four weeks (no = 0, yes = 1), violence approval, inmate–staff relationship, and facility number. Facility 1, which comprised the highest prevalence rate of physical violence, was chosen to serve as the reference category.

We entered all importation and deprivation variables considered in this study into the logistic regression analyses of prison violence simultaneously. Collinearity diagnostics of the predictors showed that none of the variance inflation factors exceeded a value of 4.5 and all tolerance values were greater than 0.2. Based on these statistics and the correlation coefficients (see Table 2) it can be assumed that the results of the regression analyses have not been affected by multicollinearity (Menard, 1995; Myers, 1990). With regard to the regression in which physical prison violence served as the outcome variable, the Hosmer–Lemeshow test produced a statistically non-significant result (χ²(8) = 12.66, p > .05), indicating a good fit of the model. R² indices are 0.24 for Cox–Snell and 0.33 for Nagelkerke (see Table 4). Inmate–staff relationship (odds ratio (OR) = 0.68, p < .01), violence approval (OR = 2.04, p < .001), and drug use (OR = 4.24, p < .001) significantly predicted involvement in physical prison violence. All other variables were not significantly related to the perpetration of physical inmate-on-inmate violence.

The predictors for sexual violence were coded in the same way as in the case of physical violence. The Hosmer–Lemeshow test yielded a non-significant result, indicating that the model fitted the data fairly well (χ²(8) = 14.98, p > .05). R² indices for the model were 0.07 (Cox–Snell) and 0.27 (Nagelkerke). As can be seen from Table 4, the only variables significantly related to sexual violence in prison are drug use during imprisonment (OR = 10.21, p < .001) and serving one’s sentence in Facility 4 (OR = 3.54, p < .05). None of the other importation and deprivation variables significantly predicted the perpetration of sexual violence.

Discussion

The present study examined predictors of physical and sexual prison violence among inmates of five YOIs in Germany, considering both importation and deprivation variables. A total of 865 male offenders participated in the study.

To implement effective means to reduce prison violence, one needs to know which factors reduce or increase the risk of perpetration. Therefore, binary logistic regression analyses were performed to examine which variables significantly predict physical and sexual inmate-on-inmate violence. Differences in perpetration rates between the facilities were found to be largely explainable by other importation and deprivation variables. Drug use during imprisonment was found to be a particularly strong predictor of both physical and sexual violence. Violent beliefs and the relationship of staff and inmates were also found to be related to physical prison violence. Noteworthy, sexual violence as opposed to physical violence was found to be significantly influenced by the facility where the young offenders are imprisoned. This suggests that variables not examined in this study, such as those concerning managerial practices, as well as the prison environment, have an effect at least on sexual inmate-on-inmate violence. However, further research is needed to examine what aspects of a prison in particular impact sexual violence.

In summary, both importation and deprivation factors were found to be significantly related to inmate-on-inmate violence. Additionally, it has to be noted that the deprivation variables that we examined, namely inmate–staff relationship and facility, are also linked to the respective prison’s management practices. It can thus be assumed that a combination of the importation, deprivation, and management models is most fruitful in predicting violent misconduct (cf. Huebner, 2003).

Inconsistent with many previous studies (e.g. Cunningham and Sorensen, 2007; DeLisi et al., 2004; Harer and Steffensmeier, 1996), age and migration background were not found to be significantly related to the perpetration of prison violence. With regard to age, this finding might be due to the fact that the inmates in the present sample were only 14–25 years old, whereas other researchers used samples with a considerably greater age range (e.g. Kuanliang et al., 2008: 13–40+ years; Lahm, 2008: 18–71 years). Regarding the weak association of migration background and prison violence, the result of this research is in line with some previous studies that reported non-significant associations of ethnicity and inmate-on-inmate violence when controlling for other variables, for example, prior gang involvement and criminal history (MacDonald, 1999; Sorensen and Cunningham, 2007). Thus, it seems that the effects of ethnicity or migration background on the perpetration of prison violence can be explained by other, probably more influential, factors.

A number of limitations have to be considered when evaluating the findings of the present study. Firstly, we only considered very few deprivation factors, namely inmate–staff relationship and the facility where the young offenders served their sentence. Future research should examine if differences in prison violence can be found on a smaller level, for example, between different wards within a single prison. Additionally, future studies should specifically evaluate the effects of further deprivation variables, such as prison crowding and security measures, on prevalence rates of violent misconduct among young offenders.

A further limitation of the current study is the fact that pre-prison drug use was not directly measured. Instead, drug use during imprisonment was treated as a proxy for prior substance abuse, because previous studies have shown that drug use prior to confinement is strongly related to substance abuse in prison (e.g. Gillespie, 2005; Jiang, 2005). An interesting endeavour for future research might be to examine to what extent the effect of drug use during imprisonment found in the present study is in fact a mere continuation of pre-prison drug use or a way of coping with the “strains” of imprisonment.

Another limitation applies to the sample that we used in this study which consisted of young males only. Research on the levels of inmate-on-inmate violence as well as its predictors among female offenders is clearly warranted, especially since previous research has found that the variables associated with violent misconduct are different for males and females (Berg and DeLisi, 2006).

The current study provided some valuable insights into the factors that significantly predict inmate-on-inmate violence. There are, however, a number of questions that remain open and should be addressed in future studies. We were able to measure a number of different importation and deprivation factors and their effects on prison violence, but our study cannot provide information on potential underlying mechanisms or mediating effects. For example, we found that drug use in prison significantly predicts physical as well as sexual inmate-on-inmate violence. However, we do not know if drug use leads to violence by reducing an inmate’s inhibition or if it rather reflects a general disposition to misconduct. Violence against fellow inmates could also be used as punishment for those who do not share their drugs (cf. Mjåland, 2014; for a framework of the drugs/violence nexus, see Goldstein, 1985). Additional research is needed to assess the causal link between drug use and violence in prison.

The results of our research point to some potential means to prevent prison violence. Previous research has shown that quality of life in prison and compliance are strongly dependent on the relationship between staff and inmates (Cesaroni and Peterson-Badali, 2010; Crewe et al., 2011; Liebling, 2008; Van der Laan and Eichelsheim, 2013). In accordance with other studies (Harer and Steffensmeier, 1996; Molleman and Leeuw, 2012), the results of the present study suggest that a negative inmate–staff relationship is also related to inmate-on-inmate violence. Therefore, carefully selecting prison staff by specifically taking into account the applicants’ attitude towards working with inmates could improve the overall staff–inmate relationship and perceptions of fair treatment and thus reduce violent misconduct. Further training for prison staff could be used to emphasize the importance of a good inmate–staff relationship in preventing prison violence. Additionally, an institutional culture should be implemented that relies on the fair treatment of prisoners and remunerative rather than coercive control mechanisms (cf. Huebner, 2003).

The finding that drug use in custody is a strong predictor of both physical and sexual violence is also important with regard to potential means of prevention. Intervention programmes for inmates who are drug users or addicts could, for example, incorporate modules geared to reduce aggression and violence potential. Targeting prisoners’ violent beliefs and attitudes in general could be a promising means to reduce violence in prison, since a significant association of violence approval and inmate-on-inmate violence has been found in this study. If one assumes that drug use itself fosters violent misconduct, rather than just representing a general underlying tendency to break rules, stricter controls and additional measures to prevent the smuggling of drugs into prison could also help to reduce the prevalence of violence among inmates.

Footnotes

Appendix 1

Table 5.

Items used to measure perpetration of physical and sexual prison violence (in part from the DIPC-SCALED-r.

Physical violence	Sexual violence
I have deliberately pushed fellow prisoners.	I have forced fellow prisoners to satisfy me with their mouth.
I have hit or kicked another prisoner.	I have forced fellow prisoners to have sexual/anal intercourse.
I have tortured fellow prisoners.
I have beaten other prisoners with an object.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

References

Baier

Bergmann

(2013) Gewalt im strafvollzug – ergebnisse einer befragung in fünf bundesländern [Violence in custody – Findings from a survey in five federal states]. Forum Strafvollzug 62: 76–83.

Berg

DeLisi

(2006) The correctional melting pot: Race, ethnicity, citizenship, and prison violence. Journal of Criminal Justice 34: 631–642. DOI: 10.1016/j.jcrimjus.2006.09.016.

Camp

Gaes

(2005) Criminogenic effects of the prison environment on inmate behavior: Some experimental evidence. Crime and Delinquency 51(3): 425–442. DOI: 10.1177/0011128704271471.

Cesaroni

Peterson-Badali

(2010) Understanding the adjustment of incarcerated young offenders: A Canadian example. Youth Justice 10(2): 107–125. DOI: 10.1177/1473225410369290.

Crewe

Liebling

Hulley

(2011) Staff culture, use of authority and prisoner quality of life in public and private sector prisons. Australian & New Zealand Journal of Criminology 44(1): 94–115. DOI: 10.1177/0004865810392681.

Crewe

Liebling

Hulley

(2015) Staff-prisoner relationships, staff professionalism, and the use of authority in public- and private-sector prisons. Law & Social Inquiry 40(2): 309–344. DOI: 10.1111/lsi.12093.

Cunningham

Sorensen

(2007) Predictive factors for violent misconduct in close custody. The Prison Journal 87(2): 241–253. DOI: 10.1177/0032885507303752.

Davidson-Arad

(2005) Observed violence, abuse, and risk behaviors in juvenile correctional facilities: Comparison of inmate and staff reports. Children and Youth Services Review 27: 547–559. DOI: 10.1016/j.childyouth.2004.11.013.

Day

Brauer

Butler

(2015) Coercion and social support behind bars: Testing an integrated theory of misconduct and resistance in U.S. prisons. Criminal Justice and Behavior 42(2): 133–155. DOI: 10.1177/0093854814546352.

10.

DeLisi

Berg

Hochstetler

(2004) Gang members, career criminals and prison violence: Further specification of the importation model of inmate behavior. Criminal Justice Studies 17(4): 369–383. DOI: 10.1080/1478601042000314883.

11.

DiIulio

(1987) Governing Prisons: A Comparative Study of Correctional Management. New York: Free Press.

12.

Ernst

(2008) Zum Ausmaß der Gewalt in deutschen Justizvollzugsanstalten: Kernbefunde einer Täter-Opfer-Befragung [On the extent of violence in German prisons: Core findings of an offender-victim-survey]. Bewährungshilfe 55: 357–372.

13.

Flanagan

(1983) Correlates of institutional misconduct among state prisoners. Criminology 21(1): 29–39. DOI: 10.1111/j.1745–9125.1983.tb00249.x.

14.

Franklin

Pratt

(2006) Examining the empirical relationship between prison crowding and inmate misconduct: A meta-analysis of conflicting research results. Journal of Criminal Justice 34: 401–412. DOI: 10.1016/j.jcrimjus.2006.05.006.

15.

Gaes

McGuir

(1985) Prison violence: The contribution of crowding versus other determinants of prison assault rates. Journal of Research in Crime and Delinquency 22(1): 41–65. DOI: 10.1177/0022427885022001003.

16.

Gillespie

(2005) A multilevel model of drug abuse inside prison. The Prison Journal 85(2): 223–246. DOI: 10.1177/0032885505277002.

17.

Goldstein

(1985) The drugs/violence nexus: A tripartite conceptual framework. Journal of Drug Issues 15(4): 493–506. DOI: 10.1177/002204268501500406.

18.

Griffin

Hepburn

(2006) The effect of gang affiliation on violent misconduct among inmates during the early years of confinement. Criminal Justice and Behavior 33(4): 419–448. DOI: 10.1177/0093854806288038.

19.

Harer

Steffensmeier

(1996) Race and prison violence. Criminology 34(3): 323–355. DOI: 10.1111/j.1745–9125.1996.tb01210.x.

20.

Hewitt

Poole

Regoli

(1984) Self-reported and observed rule-breaking in prison: A look at disciplinary response. Journal of Criminal Justice 12: 437–447.

21.

Hochstetler

DeLisi

(2005) Importation, deprivation, and varieties of serving time: An integrated-lifestyle-exposure model of prison offending. Journal of Criminal Justice 33: 257–266. DOI: 10.1016/j.jcrimjus.2005.02.005.

22.

Huebner

(2003) Administrative determinants of inmate violence: A multilevel analysis. Journal of Criminal Justice 31: 107–117. DOI: 10.1016/S0047–2352(02)00218–0.

23.

Ireland

(1999) Bullying behaviors among male and female prisoners: A study of adult and young offenders. Aggressive Behavior 25: 161–178. DOI: 10.1002/(SICI)1098–2337(1999)25:3<161::AID-AB1>3.0.CO;2-#.

24.

Irwin

Cressey

(1962) Thieves, convicts and the inmate culture. Social Problems 10(2): 142–155. DOI: 10.2307/799047.

25.

Jiang

(2005) Impact of drug use on inmate misconduct: A multilevel analysis. Journal of Criminal Justice 33(2): 153–163. DOI: 10.1016/j.jcrimjus.2004.12.007.

26.

Jiang

Fisher-Giorlando

(2002) Inmate misconduct: A test of the deprivation, importation, and situational models. The Prison Journal 82(3): 335–358. DOI: 10.1177/003288550208200303.

27.

Keene

(1997) Drug use among prisoners before, during and after custody. Addiction Research 4: 343–353. DOI: 10.3109/16066359709002968.

28.

Kuanliang

Sorensen

Cunningham

(2008) Juvenile inmates in an adult prison system: Rates of disciplinary misconduct and violence. Criminal Justice and Behavior 35(9): 1186–1201. DOI: 10.1177/0093854808322744.

29.

Kury

Smartt

(2002) Prisoner-on-prisoner violence: Victimization of young offenders in prison. Some German findings. Criminal Justice 2(4): 411–437. DOI: 10.1177/17488958020020040301.

30.

Lahm

(2008) Inmate-on-inmate assault: A multilevel examination of prison violence. Criminal Justice and Behavior 35(1): 120–137. DOI: 10.1177/0093854807308730.

31.

Lahm

(2009) Physical and property victimization behind bars: A multilevel examination. International Journal of Offender Therapy and Comparative Criminology 53(3): 348–365. DOI: 10.1177/0306624X08316504.

32.

Liebling

(2008) Incentives and earned privileges revisited: Fairness, discretion, and the quality of prison life. Journal of Scandinavian Studies in Criminology and Crime Prevention 9: 25–41. DOI: 10.1080/14043850802450773.

33.

Liebling

(2011) Distinctions and distinctiveness in the work of prison officers: Legitimacy and authority revisited. European Journal of Criminology 8(6): 484–499. DOI: 10.1177/1477370811413807.

34.

Liebling

; assisted by Arnold

(2004) Prisons and their Moral Performance. Oxford: Clarendon Press.

35.

MacDonald

(1999) Violence and drug use in juvenile institutions. Journal of Criminal Justice 27(1): 33–44. DOI: 10.1016/S0047–2352(98)00033–6.

36.

Menard

(1995) Applied Logistic Regression Analysis. Thousand Oaks, CA: SAGE, series/no. 07-106.

37.

Mills

Kroner

Forth

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

38.

Mills

Kroner

Weekes

(1998) Measuring alcohol abuse and the incidence of serious misconduct in violent offenders. The Prison Journal 78(1): 45–54. DOI: 10.1177/0032855598078001004.

39.

Mjåland

(2014) ‘A culture of sharing’: Drug exchange in a Norwegian prison. Punishment & Society 16(3): 336–352. DOI: 10.1177/1462474514527149.

40.

Moffitt

(1993) ‘Life-course-persistent’ and ‘adolescence-limited’ antisocial behavior: A developmental taxonomy. Psychological Review 100: 674–701. DOI: 10.1037/0033–295X.100.4.674.

41.

Molleman

Leeuw

(2012) The influence of prison staff on inmate conditions: A multilevel approach to staff and inmate surveys. European Journal on Criminal Policy and Research 18(2): 217–233. DOI: 10.1007/s10610–011–9158–7.

42.

Myers

(1990) Classical and Modern Regression with Applications. 2nd ed. Boston, MA: Duxbury.

43.

Neubacher

(2014) Gewalt im Jugendstrafvollzug – Ein Überblick über Ergebnisse des Kölner Forschungsprojekts [Violence in youth custody – An overview of the results of the Cologne research project]. Forum Strafvollzug 63: 320–326.

44.

Power

Dyson

Wozniak

(1997) Bullying among Scottish young offenders: Inmates’ self-reported attitudes and behaviour. Journal of Community and Applied Social Psychology 7: 209–218. DOI: 10.1002/(SICI)1099–1298(199706)7:3<209::AID-CASP419>3.0.CO;2–3.

45.

Reisig

(1998) Rates of disorder in higher-custody state prisons: A comparative analysis of managerial practices. Crime & Delinquency 44(2): 229–244. DOI: 10.1177/0011128798044002003.

46.

Sampson

Laub

(2003) Life-course desisters? Trajectories of crime among delinquent boys followed to age 70. Criminology 41(3): 555–592. DOI: 10.1111/j.1745–9125.2003.tb00997.x.

47.

Sorensen

Cunningham

(2007) Operationalizing risk: The influence of measurement choice on the prevalence and correlates of prison violence among incarcerated murderers. Journal of Criminal Justice 35: 546–555. DOI: 10.1016/j.jcrimjus.2007.07.007.

48.

Sorensen

Cunningham

(2010) Conviction offense and prison violence: A comparative study of murderers and other offenders. Crime & Delinquency 56(1): 103–125. DOI: 10.1177/0011128707307175.

49.

Sparks

Bottoms

(1995) Legitimacy and order in prisons. The British Journal of Sociology 46(1): 45–62. DOI: 10.2307/591622.

50.

Sykes

(1958) The Society of Captives. Princeton, NJ: Princeton University Press.

51.

Tasca

Griffin

Rodriguez

(2010) The effect of importation and deprivation factors on violent misconduct: An examination of Black and Latino youth in prison. Youth Violence and Juvenile Justice 8(3): 234–249. DOI: 10.1177/1541204010366619.

52.

Thomas

(1977) Theoretical perspectives on prisonization: A comparison of the importation and deprivation models. Journal of Criminal Law and Criminology 68: 135–145. DOI: 10.2307/1142482.

53.

Thomas

Cage

(1977) Correlates of prison drug use: An evaluation of two conceptual models. Criminology 15: 193–210. DOI: 10.1111/j.1745–9125.1977.tb00061.x.

54.

Trulson

Marquart

Mullings

. (2005) In between adolescence and adulthood: Recidivism outcomes of a cohort of state delinquents. Youth Violence and Juvenile Justice 3(4): 335–387. DOI: 10.1177/1541204005278802.

55.

Van der Laan

Eichelsheim

(2013) Juvenile adaptation to imprisonment: Feelings of safety, autonomy and well-being, and behaviour in prison. European Journal of Criminology 10(4): 424–443. DOI: 10.1177/1477370812473530.

56.

Wirth

(2007) Gewalt unter gefangenen: Kernbefunde einer empirischen studie im strafvollzug des Landes Nordrhein-Westfalen [Violence among inmates: Core findings of an empirical study of prisons in North Rhine-Westphalia]. Bewährungshilfe 54: 185–206.

57.

Wolter

Boxberg

(2016) The perception of imprisonment and its effect on inmate violence. In: Reeves

(ed) Experiencing Imprisonment: Research on the Experience of living and working in Carceral Institutions. London: Routledge, 156–175.