Differences between boys and girls in perceived group climate in residential youth care

Abstract The aim of the present study was to examine differences in perceived living group climate between boys and girls in a sample of 344 youth (68.6% male, M age = 16, SD = 1.58) receiving residential youth care in the Netherlands. Participants filled out self-report measures on living group climate. Results of multilevel regression models indicated that girls in non-correctional facilities experienced living group climate most positively, and girls in correctional facilities experienced living group climate most negatively compared to girls and boys in non-correctional facilities. We conclude that residential treatment settings should adapt gender-responsive approaches to address specific needs of girls, specifically in secure residential care. Future studies should focus on specific needs of girls to advance knowledge on how they can benefit optimally from their stay in residential care to facilitate gender-specific programming in residential youth care.


Introduction
Adolescents with severe behavioral problems or who commit crimes are treated in (secure) residential youth care facilities. In general, their problems are more complex and severe compared to youth in nonresidential facilities (Leloux-Opmeer, Kuiper, Swaab, & Scholte, 2016;Ter Beek, Van der Rijken, Kuiper, Hendriks, & Stams, 2018;Vermaes & Nijhof, 2014). Residential care and treatment in the Netherlands are delivered in living groups of typically 8-10 adolescents, where adolescents are supervised by two or more trained group workers. In correctional facilities, rules and regulations are mostly restrictive, and contact with peers is supervised by staff (De Valk, Kuiper, Van der Helm, Maas, & Stams, 2016).
In the Netherlands in the period 2010-2012 1 , it is estimated that 90 per 10.000 minors were placed out of their homes each year, ca. 45,000 in total. Over the same period, the lowest number of young people in Europe were placed out of their homes in Italy − 38 per 10,000 minorswhile the highest numbers were to be found in Denmark 120 per 10,000 minors. This means that the Dutch figures are slightly above the European average of the Netherlands (Harder, Knorth & Kuiper, 2020, p. 16).
In the year 2013 2 , 11,540 of these children and adolescents received residential youth care (Dutch Central Bureau of Statistics, 2020). Most youths are likely to be referred to open (non-correctional) residential youth care facilities, where as many girls as boys live. Of this group (in 2013) 1,818 adolescents (Jeugdzorg Nederland, 2017) between 12 and 18 years of age have been placed in one of twelve secure (non-correctional) youth care facilities (called Youthcare Plus ) to protect them from others who pose a threat to their development, to prevent them from self-harm or refusal of necessary care, but most of all to provide intensive 24-hour care and effective treatment for complex problems ( Van der Helm, Kuiper, & Stams, 2018). There is always a judicial authorization for the placement. Of these adolescents in Youthcare Plus 43% were female and 57% male (Jeugdzorg Nederland, 2017). In the year 2013 on top of the adolescents in residential youth care 1,180 (5,4% girls in a girls-only group) were placed in youth prisons (WODC, 2020). In those cases, there was a form of punishment related to a criminal offence. The ultimate aim of residential treatment for youth is to learn to get along with others, to (re)start schooling, to develop prosocial attitudes, to reduce internalizing and externalizing problem behavior, delinquency, substance use, and prevent re-victimization ( Van der Helm et al., 2018;Vermaes & Nijhof, 2014).
In order to achieve youths' treatment goals, a positive living group climate is required, as the living group is the primary social environment for adolescents receiving residential care (Eltink, 2020;Leipoldt, Harder, Kayed, Grietens, & Rimehaug, 2019;Van der Helm et al., 2018). Knowledge on how girls and boys differ in their perception of living group climate is important to better understand how youth in residential youth care can benefit from their treatment in order to facilitate gender-specific programming in residential care. Notably, the perception of living group climate by girls is an important although understudied subject in literature. Therefore, the aim of the present study was to examine differences between boys and girls in perceived living group climate in secure (correctional and noncorrectional) facilities.

Gender responsive treatment in residential youth care
Effective treatment is based on the RNR principles (Andrews & Bonta, 2010), which indicate that treatment should be tailored to the individual needs of the person: the intensity of treatment should be in line with the risk for recidivism or revictimization, should fit the criminogenic or development threatening needs, and should be in line with the motivation and capabilities of the individual. Recent studies have emphasized the need for gender responsive treatment of youth in residential care based on differences in psychological development of boys and girls as well as differences in exposure to risk factors, pathways to crime, and needs (Anderson, Hoskins, & Rubino, 2019;Granski et al., 2020;Hubbard & Matthews, 2008;Lanctôt, 2018;Piller, Gibly, & Peled 2019;Walker, Bishop, Nurius, & Logan-Greene, 2016). For example, boys tend to display more aggressive and delinquent behavior than girls, whereas girls display more internalizing behavior and are more at risk for exposure to adverse childhood experiences (ACEs), such as parental neglect, domestic violence, and sexual abuse (Asscher, Van der Put, & Stams, 2015;Assink et al., 2019;Biswas & Vaughn, 2011;Chaplo, Kerig, Modrowski, & Bennett, 2017;Dirkse, Eichelsheim, Asscher, & Van der Laan, 2018;Leve, Chamberlain, & Kim, 2015;Van Damme, Colins, De Maeyer, Vermeiren, & Vanderplasschen, 2015).
A recent meta-analysis by Granski et al. (2020) on program characteristics for youth with disruptive behavior problems demonstrated that boys may benefit more from interventions targeting disruptive behavior problems than girls. This finding could imply that gender responsive treatment for disruptive behavior problems is necessary to allow girls to benefit equally from treatment as boys. The authors suggest that the delivery of trauma-informed approaches to address experiences of trauma and victimization in girls as well as a relational approach, focusing on the girls' family and peer group, may be more beneficial for girls.
During the past decade, research has focused on treatment principles and implementation of trauma-informed care in residential youth care facilities aimed at preventing re-victimization and traumatization, particularly through responsive and non-coercive staff-client interactions by refraining from restrictive measures, such as seclusion and restraint (Bryson et al., 2017;Ford et al., 2012;Ford & Blaustein, 2013;Hodgdon, Kinniburgh, Gabowitz, Blaustein, & Spinazzola, 2013;Roy et al., 2019). Also, recent studies have specifically focused on relational approaches targeting girls in residential care (Lanctôt, 2018;Lanctôt, Lemieux, & Mathys, 2016). While research on trauma-informed care and gender-responsive treatment is steadily growing, the effective implementation of these practices remains limited (Anderson et al., 2019;Bryson et al, 2017;Lanctôt, 2018, Smith, 2017. In addition to the development of promising interventions based on trauma-informed or relational based principles to address girls' needs in residential care, the adaptation of gender-responsive strategies may also be effective at the living group in which youth stay during their treatment.

Living group climate
During their stay in residential care it is important for youth to develop prosocial attitudes and prosocial behavior. The living group in which youth reside is the primary social environment for adolescents in residential care. The quality of this environment can be described in terms of living group climate. Living group climate can be defined as 'the quality of the social and physical environment in terms of the provision of sufficient and necessary conditions for physical and mental health, well-being, contact, and personal growth of the residents, with respect for their human dignity and human rights, as well as (if not restricted by judicial measures) their personal autonomy, aimed at recovery and successful participation in society' (Stams & Van der Helm, 2017, p. 4).
Living group climate may be relatively 'open', providing a structured therapeutic environment, with a positive atmosphere where the adolescents respect each other and feel safe, and where group social workers support their needs, and support cohesion and active participation. On the other hand, group climate may be negative and repressive, with a lack of mutual respect and safety among the adolescents and between adolescents and group workers. An open group climate is associated with well-being, increased treatment motivation, active coping strategies, and treatment satisfaction (Leipoldt et al.,  In recent years, studies have focused on girls' perceptions of living group climate in residential youth care. Several studies have found relational factors, such as relationships with peers and relationships with staff, to be important aspects of group climate for girls (Cantora, 2014;Kerig & Schindler, 2013;Mathys, Lanctôt, & Touchette, 2013). Previous studies have demonstrated that girls are more orientated towards social interaction and social support seeking (Miller, Leve & Kerig, 2012;Miller, Winn, Taylor, & Wiki, 2012;Taylor et al., 2000). Lanctôt, Lemieux and Mathys (2016) explored girls' perceptions of group climate by means of latent class analysis, revealing that girls differ in their perceptions of group climate in terms of their sense of safety among their peers at the group, their connection with their peers, and with staff. Also, girls who displayed more complex problems at admissionparticularly trauma-related problems -had more negative perceptions of group climate (Lanctôt et al., 2016). A qualitative study by De Valk, Kuiper, Van der Helm, Maas and Stams (2017) revealed differences between boys and girls in how youth perceived autonomy and meaningfulness. Girls experienced more repression when they were not granted opportunities for personal development related to internally motivated goals, whereas boys mainly focused on meeting externally motivated goals. For example, they wanted to meet the expectations of staff in order to be 'released'. Studies focusing on the working alliancethe collaborative relationshipbetween girls and staff in residential care found that staff found it more difficult to establish a working alliance with girls than with boys due to gender-specific extreme problem behavior, such as auto-mutilation (Lanctôt, Ayotte, Turcotte, & Besnard, 2012). Additionally, Ayotte Lanctôt and Tourigny (2015) found that girls with more severe problem behavior had weaker working alliances with staff.

Present study
The aim of the present study was to examine differences between boys and girls in perceived living group climate in correctional and noncorrectional facilities. To the best of our knowledge, prior studies on perception of group climate in residential youth care have not examined differences between boys and girls, taking into account the type of facility (non-correctional versus correctional facilities). Given the higher degree of coercion and restrictiveness and more severe problem behavior of youth in correctional facilities (Vermaes & Nijhof, 2014), we expected youth in these facilities to experience a less positive living group climate compared to youth in non-correctional facilities.

Participants
The total (convenience) sample consisted of 344 adolescents (n = 236 boys, n = 108 girls) receiving residential care in the Netherlands. The mean age of the respondents was 16 years (SD = 1.58; range 7-24) and 77.9% had the Dutch nationality. Of the total sample, 76 youth (n = 40 boys, n = 36 girls) were treated in non-correctional centers (5 different open facilities) and 268 youth (n = 196 boys, n = 72 girls) were treated in correctional centers (6 different secure care facilities and 4 different youth prisons). The living groups at which youth were placed consisted of boys-only, girls-only, and mixed groups. Of the total sample, 179 boys resided in boys-only groups, 62 girls resided in girls-only groups, and 96 youth were resided in mixed groups (n = 53 boys, n = 43 girls). Youth were referred to secure non-correctional centers because of severe behavior problems or being a danger to themselves or others.

Procedure
The data were collected in 2013. All adolescents participated voluntarily, signed an informed consent declaration, and were told that their answers would be treated confidentially and anonymously, and would be accessed only by the researchers. As a token of gratitude for their participation, participants received a telephone card or a small gift with a maximum value of €5.50. The names of the respondents on the questionnaires were deleted and substituted with a code number for use in SPSS data-analysis software. In order to protect the privacy of the adolescents, researchers had no access to their full names. Specially trained graduate students of social work administered questionnaires during a group session on the residential group. The questionnaires were completed two to three weeks after the arrival of the adolescent at the centers. This choice (cross-sectional) has been made because the duration of stay of the youngsters varies considerably from for example, 6 weeks in a juvenile prison to 2 years in an open residential facility. The answering rate was approximately 82%; reasons for not participating were: absence due to going to court or leave (12%); not interested or angry (6%).

Measures
Living group climate. Living group climate was assessed with the Prison Group Climate Inventory (PGCI, Van der Helm et al., 2011). This self-report measure consists of 36 items rated on a five point Likert type scale, ranging from 1, (I do not agree), to 5 (I totally agree). Each item belongs to one of the four scales for group climate: Support, Growth, Repression and Atmosphere. The Support scale (12 items) assesses perceived support from group workers, in particularly the responsiveness of group workers. Listening to youth, taking their complaints seriously, and respect and trust are important characteristics of support. An example item is: 'Group workers treat me with respect'. The Growth scale (8 items) assesses youth's perceptions of learning and hope for the future during their stay in the center. An example item is: 'I learn the right things here'. The Repression scale (9 items) assesses perceptions of strictness and control as well as unfair and haphazard rules. An example item is: 'You have to ask permission for everything here'. The Atmosphere scale (7 items) assesses the social interaction among youth in terms of mutual trust, their feelings of safety at the group, and how youth perceive the physical environment at the group, such as daylight and fresh air at the group. An example item is: 'We trust each other here'. The internal consistency reliability of the scales was good in boys and girls for the scales Support (α = 0.90, α = 0.87, respectively), Growth (α = 0.87, α = 0.88, respectively), Repression (α = 0.75, α = 0.72, respectively), and Atmosphere (α = 0.78, α = 0.83, respectively). Higher mean scores for the scales Support, Growth, and Atmosphere are indicative of higher levels of support from group workers, more possibilities for growth, and a more positive atmosphere as perceived by youth. Also, a higher mean score for the scale Repression is indicative of more repressive behavior of staff as perceived by youth.

Statistical analyses
To examine differences in living group climate, a multilevel approach was used, given the nested nature of the data (youth are nested in living groups). Hierarchical Linear Modeling (HLM) was used to account for violation of the independence assumption of regression. HLM allows for examination of how variation in the dependent variable is attributed to differences within-group (i.e., individual level) or betweengroup (i.e., living group level). The analyses were conducted using the 'lme4 ′ package (Bates, Maechler, Bolker, & Walker, 2015) in the R environment. The 'lmerTest' package (Kuznetsova, Brockhoff, & Christensen, 2015) was used for the calculation of p-values, which uses the Satterthwaite approximation procedure for calculating degrees of freedom.
Four models were fit for each dependent variable. First, a random intercept-only model (null model) was fitted without predictors to estimate the Level-2 variance and ICC (Intraclass Correlation Coefficient) for the dependent variable. When significant Level-2 variance is demonstrated, multilevel analysis is warranted. Subsequently, three multilevel models were fitted. The first model included only main effects of Level-1 predictors (Gender and Age). A second model included Level-2 predictors (Type of facility). A final model included a cross-level interaction between Gender and Type of facility. The fit of the models was compared using likelihood-ratio tests. Models were compared after adding the cross-level interaction term. Parameter estimates and statistical tests on individual terms are reported for the basic model and the final model. Table 1 shows the descriptive information of all study variables for boys and girls in non-correctional and correctional residential care facilities. Mean scores of Support, Growth, and Atmosphere A series of hierarchical linear models was conducted to examine the relation between gender, type of facility, and living group climate. First, four different random intercept-only models were conducted to establish whether there was significant variance at Level-2 for the variables Support, Growth, Repression, and Atmosphere. Results indicated that there was significant Level-2 variance for Support (ICC = 0.16, Wald z = 2.68, p = .007), Growth (ICC = 0.16, Wald z = 2.55, p = .011) Repression (ICC = 0.22, Wald z = 3.04, p = .002), and Atmosphere (ICC = 0.16, Wald z = 2.50, p = .012).

Results
Second, for each dependent variable, multilevel models were specified in which several main effects for Level-1 (Gender and Age) and Level-2 (Type of facility) were included, and an interaction effect between Gender and Type of facility.
Results of the HLM models with only main effects (basic model) and interaction effects (final model), as well as comparisons between models are presented in Tables 2-5. Results indicated a significant cross-level interaction effect between Gender and Type of facility for Support (β = -0.75), Growth (β = -0.76), and Atmosphere (β = -0.87), which indicated that girls in correctional facilities experienced lower levels of support, growth, and atmosphere compared to boys. Results showed no significant interaction effect between gender and type of facility for Repression. However, a positive main effect was found for Type of facility (β = 0.85), indicating that youth in correctional facilities experienced more repression compared to youth in non-correctional facilities.

Discussion
The aim of the present study was to examine differences in perceived group climate between boys and girls in non-correctional and correctional residential care facilities. Findings indicated that girls in noncorrectional facilities experienced aspects of living group climate most positively (more support, more possibilities for growth, and a more positive atmosphere), and girls in correctional facilities experienced aspects of living group climate most negatively compared to girls and boys in non-correctional facilities. The differences in perceived group climate between youth in non-correctional and correctional facilities were largely in line with our expectations. According to literature, these results could be explained by the higher degree of coercion and restrictiveness as well as by more severe problem behavior of youth in correctional facilities (Vermaes & Nijhof, 2014). For girls, this possibly touches on giving meaning to their stay because they experience less possibilities for personal growth (De Valk et al., 2017). In addition, girls may benefit less from interventions targeting disruptive behavior problems (Granski et al, 2020).
Findings also revealed that girls in non-correctional facilities indeed reported more support from group workers, whereas girls in correctional facilities experienced less support from group workers. These findings could also be explained by the complex and severe problem behavior of girls in correctional facilities. Most girls in correctional facilities have a history of trauma and are often poly-victimized (Kerig & Becker, 2012). Lanctôt (2020) found that high levels of childhood trauma in girls are related to maladaptive cognitive schemas, specifically perceptions of disconnection and rejection, which is negatively related to girls' perceived social support during their stay in residential care. Several other studies found that severe problem behavior in girls is related to a poor working alliance with staff (Ayotte et al., 2015(Ayotte et al., , 2017Lanctôt et al., 2012).
Further, findings revealed that girls in correctional facilities   experienced a less positive atmosphere at the living group. Prosocial and meaningful relationships are necessary to feel 'safe and connected', and prior research suggests that this is specifically the case for girls (Lanctôt et al., 2016). This means that staff in correctional facilities should take an active role in stimulating positive interactions among youth (Sonderman et al., 2020). Further, by creating a positive atmosphere at the living group, aggressive behavior and negative group dynamics may be reduced or prevented. This is important not only to create a positive living group climate during youth's stay at the facility, but also because evidence suggests that aggressive behavior persists in adulthood (Cleverley, Szatmari, Vaillancourt, Boyle, & Lipman, 2012;Krabbendam et al., 2014;Teplin, Welty, Abram, Dulcan & Washburn, 2012).

Limitations
Several limitations to the present study need to be acknowledged. Firstly, the present study used cross-sectional data and the measurements were carried out at the beginning (first weeks) of the youths' stay in the facility. A longitudinal design would have provided a better insight in how youth experience living group climate throughout their stay. A second limitation is that we did not assess problem behavior of youth. Severity of problem behavior of youth, specifically traumarelated symptoms in girls, is important to consider when measuring group climate. Thirdly, the living group climate instrument used in the present study not only measures atmosphere at the living group in terms of social interactions between youth, containing items referring to feelings of safety and mutual trust, but also as the quality of the physical environment, which somewhat obscures how youth perceive interactions with other group members. Finally, due to the limited sample size, we were not able to examine measurement invariance across boys only, mixed gender and girls only groups.

Implications for practice and Future research
A number of implications for practice and future research can be formulated based on the findings of the present study. Since the results of this study indicate that in particular girls in correctional facilities experience a negative living group climate, this subject warrant further investigation. In correctional settings, youth cannot leave the facility voluntarily, it is more difficult for them to maintain pre-existing positive relationships with their family and peers, and therefore youth have very limited choice in people to interact with. This means that most social interactions among youth and between youth and staff take place at the living group. Therefore, in the context of secure residential care it is especially important to address specific needs of girls regarding their social interactions and need for support. In this regard, Lanctôt (2018) found that girls expected more from staff and their mentor than boys. According to Lanctôt (2018), staff members must be particularly responsive to the feelings and emotions of the girls, and must demonstrate a genuine desire to help them, whereas boys may need more supervision and guidance for dealing with rules and restrictions.
Future research should focus on what knowledge and skills group workers need for working with girls in secure youth care, and whether group workers could benefit from training in specific treatment approaches that are suitable within a sociotherapeutic context, such as a trauma-informed approach. It should be recognized that traumainformed care requires a paradigm shift within any youth care facility, focusing more on understanding trauma, stress, and their impact on client-staff interactions, informed by principles of collaboration as well as shared decision making (Bryson et al., 2017;Roy, Morizot, Lamothe, & Geoffrion, 2020). This perspective is in line with principles of working on a positive living group climate, which not only requires effort and commitment from staff at the living group in terms of providing support, autonomy granting, and refraining from repressive behavior, but also a positive working climate of staff in terms of team functioning and a safe work environment Roy et al., 2020). Providing group workers tools to help youth cope during their stay and focusing on prosocial behavior, specifically by stimulating positive peer interactions and relationships with their social network may improve treatment efficacy and increase motivation for change.
Future studies should focus on perception of living group climate of youth using a longitudinal design, taking into account correlates such as problem behavior, trauma-related symptoms, aggressive behavior at the living group, as well as motivation for treatment. Also, establishing measurement invariance of group climate scales is needed to examine whether group climate constructs can be measured in a conceptually identical way across boys and girls, which allows meaningful comparisons of perception of group climate between boys and girls. Further, qualitative methods are necessary to investigate how girls and boys in secure residential care perceive living group climate.
Finally, in addition to establish a clear understanding of how boys and girls in correctional facilities differ in their criminogenic and treatment needs the heterogeneity of needs of boys and girls should be addressed by recognizing that they are not homogeneous groups (Lanctôt, 2018). A more detailed understanding of how girls and boys  differ in their perception of living group climate and acknowledging differences within girls and boys' populations are important to advance knowledge on how youth may benefit from their stay in residential care. Future qualitative research on the differences can contribute to a deeper understanding of the gender differences that emerged in this study, which knowledge can be used to inform practice in order to facilitate gender-specific programming as well as a client-centered approach in residential youth care.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.