Sexual and gender-based violence in the European asylum and reception sector: a perpetuum mobile?

Abstract

Background: Refugees, asylum seekers and undocumented migrants are at risk of sexual and gender-based violence (SGBV) and subsequent ill-health in Europe; yet, European minimum reception standards do not address SGBV. Hence, this paper explores the nature of SGBV occurring in this sector and discusses determinants for ‘Desirable Prevention’. Methods: Applying community-based participatory research, we conducted an SGBV knowledge, attitude and practice survey with residents and professionals in eight European countries. We conducted logistic regression using mixed models to analyse the data in R. Results: Of the 562 respondents, 58.3% reported cases of direct (23.3%) or peer (76.6%) victimization. Our results indicate that when men were involved, it most likely concerned sexual perpetration (adjusted odds ratio [aOR]: 4.09, confidence interval [CI]: 1.2; 13.89) and physical victimization (aOR: 2.57, CI: 1.65; 4), compared with females, who then rather perpetrated emotional violence (aOR: 1.85, CI: 1.08; 3.13) and underwent sexual victimization (aOR: 7.14, CI: 3.33; 16.67). Compared with others, asylum seekers appeared more likely to perpetrate physical (aOR 7.14, CI: 4; 12.5) and endure socio-economic violence (aOR: 10, CI: 1.37; 100), whereas professionals rather bore emotional (aOR: 2.01, CI: 0.98; 4.12) and perpetrated socio-economic violence (aOR: 25.91, CI: 13.41; 50.07). When group perpetration (aOR: 2.13, CI: 1.27; 3.58) or victimization (aOR: 1.84, CI: 1.1; 3.06) occurred, it most likely concerned socio-economic violence. Conclusion: Within the European asylum reception sector, residents and professionals of both sexes experience SGBV victimization and perpetration. Given the lack of prevention policies, our findings call for urgent Desirable Prevention programmes addressing determinants socio-ecologically.

Introduction

Sexual and gender-based violence (SGBV) is generally defined as ‘any act as well as threats of acts of physical, sexual and psychological violence that is directed against a person on the basis of her/his gender or sex, and which occurs in the family, the community, or is perpetrated or condoned by the State and/or institutions’.¹ Yet, in the context of (forced) migration, the United Nations High Commissioner on Refugees (UNHCR) applies a definition that comprises five types of violence, namely, physical, psychological, sexual, socio-economic violence and harmful cultural practices.² Given our research population, we adopt this definition of SGBV and refer to the different categories in our analysis. In addition to important adverse effects on the victim’s well-being and participation in society, SGBV may induce long-lasting ill sexual, reproductive, physical and mental health,^3–7 primarily affecting the victim, yet also potentially harmful to the victim’s peers, offspring and community.^8–11

SGBV is a global public health issue and a violation of human rights. The World Health Organization (WHO) recently stated that 25.4% of women and girls in the European WHO Region experience physical and or sexual violence by their (ex)partner and 5.2% are sexually victimized by non-partners.¹² For both sexes, one is considered vulnerable when being an adolescent of low socio-economic status;^3,13 when being an undocumented migrant (no legal residence status), an asylum seeker (temporary legal residence) or a refugee (legal residence);^14–16 and when living in shelters, rehabilitative facilities or detention.^19–20 People who were personally victimized or witnessed SGBV during childhood and those with a heightened risk perception due to victimization of linked people are prone to subsequent victimization or perpetration of SGBV at later stages of life.^{10,11,21–23} Although perpetrators of SGBV are commonly known to the victim,^3,12,24 research has shown that boys, migrants or impoverished people are equally victimized by strangers, persons in authority and those assigned to their protection.^{4,13,15,25,26}

The EU Member States received 296 700 new asylum claims in 2012, which is an increase of 7% compared with 2011;²⁷ refugee status was granted to 14% of them.²⁸ The asylum systems differ greatly from country to country, with housing facilities ranging from hangars and tents (Malta) to hotel rooms (Ireland and Belgium) and from small local reception initiatives in houses to big open accommodation or detention centres (in all research countries). A lot of European countries face difficulties in upholding the European minimum standards of reception,²⁹ which lacked SGBV prevention measures until the recast of June 2013.³⁰

Although several determinants in SGBV are known, it is unclear how SGBV in the European asylum reception facilities is linked with current violence prevention knowledge, attitudes, practices and needs of residents and staff. Hence, this paper explores the nature of violence that residents and professionals experienced within the reception facilities in the year prior to the interview and discusses which determinants are decisive for ‘Desirable Prevention’ of SGBV in these settings.

Methods

The conceptual framework comprised the socio-ecological model on health and violence,³¹ incorporating ‘Desirable Prevention’³² and community-based participatory research (CBPR). CBPR in public health focuses on inequalities and aims to improve the health and well-being of community members by integrating knowledge in action, including social and policy change.^33,34 Applying this framework, we mobilized stakeholders in the eight countries of research (Belgium, Greece, Hungary, Ireland, Malta, The Netherlands, Portugal and Spain) to participate in community advisory boards (CAB). These CAB consisted of asylum seekers and refugees, asylum reception professionals, policymakers, intermediary organizations, civil society and researchers engaged in the asylum and reception sector. The CAB participated in every decisive phase of the project.

We considered the residents and professionals as the research’s main beneficiaries and set out different inclusion criteria. For the residents this implied being member of the most numerous groups of asylum-seeking and unaccompanied minor communities. They had to be staying at, or just having left, an asylum reception facility in the country of research. The professionals had to work, or had just stopped working, there. All official and unofficial types of reception facilities were included and a geographical distribution over the country of research was sought. Subsequently, per country, we recruited one to three professionals and four to seven residents who demonstrated good social and communication skills. They completed a standardized 24-hour training course based on which they became ‘Community Researchers’ (CR) participating collegially throughout the project. Together with the CR and the CAB, we developed a knowledge, attitude and practice (KAP) survey. The questionnaire was translated and back-translated into the main languages of potential respondents, and pilot-tested with the CAB.

Fieldwork

The KAP survey was conducted between October 2009 and August 2010. We listed all actually used types of reception facilities and services per country and then randomly sampled them in six of the eight countries. Due to political constraints, we were obliged to adopt convenience sampling in The Netherlands and Spain. After having obtained the permission of the sampled facilities, we applied the inclusion criteria and then randomly sampled the respondents on their list of residents and professionals. Interviews were scheduled for a one-to-one interview with the CR at a private place in or near the centre in a language the respondent and the CR both mastered well and commonly agreed upon. Respondents were informed about the study and participation modes and guaranteed that their participation would not affect their asylum case and that analysis would be anonymous. Informed consent was obtained in writing.

The KAP questionnaire comprised three parts. First, we inquired about respondents’ knowledge and experience of violence at the reception facility. If they answered positively; respondents could describe three violence cases of the year prior to the interview. For each of them, we checked the victim’s and perpetrator’s sex, legal status, age, operation modus and relationship to each other and to the respondents. Respondents had the opportunity to disclose personal involvement both directly and indirectly. Second, we explored the respondents’ attitudes towards violence and its prevention. Third, we investigated the currently applied practices in violence prevention and response. Upon completion of the interview, the respondents were given information and referral material on health and violence in the language of the interview and a small incentive in phone credit or body care products. In Ireland no incentive was given, in line with the University College of Dublin’s ethical guidelines. Respondents were invited to participate in further project phases. The study protocol applied the WHO and UNHCR ethical and safety guidelines in researching violence, complied with the local ethical requirements and received ethical approval from the Ghent University Hospital Ethical Committee [B67020096667].

Data analysis

In the eight countries, 600 individual interviews were conducted. The CR handed the interview guide, their notes, the translation of the open questionnaire and the signed informed consent to the country coordinator. (S)he checked validity and sent it on to the international project coordinator who did a second round of validity checking, based on which interviews and informed consent were separated. In all, 562 interviews were withheld, while all 38 Spanish interviews were excluded, as their validity could not be guaranteed. The quantitative data were put into an SPSS database. We applied the Framework Analysis Technique to analyse the qualitative data, a process conducted by three researchers who eventually consented on a set of categories that were then added to the SPSS database. Eventually, R was used for analysis.

We conducted logistic regression analysis using generalized linear models and mixed models to evaluate the relationship between types of violence and specific characteristics of the victims and perpetrators. First, we built generalized linear models assuming no cluster effect. Second, we used the same generalized linear models but accounting for possible clustering at country level. Finally, we performed an analysis with mixed models. All observations in the analysis regarded cases including at least one type of violence. The outcome in all models was a binary variable corresponding to the specific presence of the following types of violence in reported cases: (i) physical, (ii) emotional, (iii) sexual and (iv) socio-economic violence. The independent variables and fixed effects in the mixed models were sex, age and legal status of both the victim and the perpetrator; whether the victim was victimized in a group or alone; and whether the perpetrator acted in a group or alone. We included the country variable as a random effect in our mixed model and as cluster variable in our second generalized linear model. Models were estimated using the following functions and packages in R 3.0.1 and R Studio 0.97.551: ‘glm' (‘stats’ package), ‘surveyglm’ (‘survey’ package) and ‘glmer’ (‘lme4' packages). Model selection was performed using Akaike’s Information Criterion (AIC), where the model with the lowest AIC value was considered the best model. In 19 of the 32 (60%) tested associations, the mixed model had lower AIC values than the generalized linear model. The results presented in the following text are based on the mixed-model results, but do not differ significantly from those obtained with generalized linear models. Adjusted odds ratios (aORs) accounting for inter-country variation are given in the regression models.

This paper focuses solely on the violent experiences that were reported in the first part of the KAP study. Duplicates of cases were deleted in the data cleaning rounds. The preliminary results were presented to all CABs and interpreted together to facilitate the development of the ‘Senperforto Frame of Reference on SGBV prevention and response’.

RESULTS

Socio-demographic profile of respondents and their experience with violence

The 562 respondents comprised 375 (66.73%) residents and 187 (33.27%) professionals. Of the 562 respondents, 234 (41.64%) did not report cases they perceived as violence, while 328 (58.36%) did. The latter described 600 different cases consisting of personal victimization in 23.67% (142) and victimization of a co-resident or professional in 76.33% (458) in the asylum setting they live/work. Table 1 gives an overview of the profile of respondents reporting no, personal or peer victimization.

Table 2 demonstrates that the reported cases consisted mostly of a combination of multiple acts of different violence types that can be categorized as physical (n = 437), emotional (n = 420), sexual (n = 62) and socio-economic violence (n = 117).

Table 3 shows that both sexes as well as both residents and professionals are at risk of victimization and perpetration within the European asylum reception sector. Yet, our results suggest that each of them is more likely to be involved in a specific type of violence and operation modus.

Table 1

Reported victimization of respondents and peers

Socio-demographic characteristics	Reported no violence		Reported being personally victimized		Reported only violence against peer		Total
	N (228)	%	N (112)	%	N (217)	%	N (557)	%
Country
Belgium	31	13.6	35	31.3	26	12.0	92	16.5
Greece	23	10.1	14	12.5	29	13.4	66	11.8
Hungary	43	18.9	18	16.1	28	12.9	89	16.0
Ireland	26	11.4	12	10.7	55	25.3	93	16.7
Malta	25	11.0	17	15.2	47	21.7	89	16.0
The Netherlands	2	0.9	12	10.7	24	11.1	38	6.8
Portugal	78	34.2	4	3.6	8	3.7	90	16.2
Sex
Female	94	41.2	50	44.6	95	44.0	239	43.0
Male	134	58.8	62	55.4	121	56.0	317	57.0
Missing					1		1
Age (years)
12–18	41	18.6	24	22.4	15	7.4	80	15.1
19–29	75	33.9	25	23.4	74	36.6	174	32.8
30–39	73	33.0	38	35.5	70	34.7	181	34.2
40–49	18	8.1	12	11.2	21	10.4	51	9.6
>50	14	6.3	8	7.5	22	10.9	44	8.3
Missing	7		5		15		27
Legal status
Asylum seeker	88	38.8	50	44.6	85	39.4	223	40.2
Humanitarian and subsidiary protection	54	23.8	9	8.0	24	11.1	87	15.7
Recognized refugee	32	14.1	6	5.4	8	3.7	46	8.3
Undocumented	9	4.0	4	3.6	9	4.2	22	4.0
National citizen	42	18.5	43	38.4	86	39.8	171	30.8
Other	2	0.9	0	0.0	4	1.9	6	1.1
Missing	1				1		2
Facility type
Detention centre	12	5.4	5	4.6	7	3.5	24	4.6
Open reception centre (incl. unaccomp minors)	142	64.3	82	75.9	146	72.7%	366	69.6
Local reception initiative	50	22.6	10	9.3	23	11.4	83	15.8
Return centre	2	0.9	4	3.7	0	0.0	6	1.1
Other (e.g. hotel, health service, … )	15	6.8	7	6.5	25	12.4	47	8.9
Missing	7		4			16		27

Table 2

Overview types of violence acts in reported cases

Types of violence acts	Respondent = victim	Peer = victim	Total acts (n = 1036)	100%
Emotional violence	87	333	420	40.54
Verbal violence	34	125	159	15.35
Humiliation	35	74	109	10.52
Threatening	16	89	105	10.13
Confinement	1	7	8	0.77
Relational violence	1	38	39	3.76
Physical violence	73	364	437	42.18
Singular non-life threatening	37	201	238	22.97
Multiple non-life threatening	12	51	63	6.08
Singular life threatening	12	81	93	8.98
Multiple life threatening	12	28	40	3.86
Killing	0	3	3	0.29
Sexual violence	13	49	62	5.98
Sexual harassment	8	23	31	2.99
Sexual abuse	2	10	12	1.16
Attempt to rape	0	1	1	0.10
Rape	0	6	6	0.58
Sexual exploitation	3	9	12	1.16
Socio-economic violence	38	79	117	11.29
Discrimination	10	31	41	3.96
Refusal of assistance	25	39	64	6.18
Social exclusion	2	7	9	0.87
Refusal of legal protection	1	2	3	0.29
Total	211	825	1036	100%

Table 3

Characteristics of perpetrators and victims per type of violence. (P-values: univariate analysis adjusted for country)

Characteristics of perpetrators and victims	Physical				Emotional				Sexual				Socio-economic
	Peer = victim		Resp = victim		Peer = victim		Resp = victim		Peer = victim		Resp = victim		Peer = victim		Resp = victim
	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95%CI)	P-value	OR (95% CI)	P-value	OR (95%CI)	P-value
Sex of the perpetrator (Ref = female)
Male	1.00 (0.57–1.74)	0.996	1 47 (0.56–3.88)	0.433	0 54 (0.32–0.93)	0.027	1 13 (0.39–3.32)	0.818	4 09 (1.20–13.89)	0.024	1 65 (0.32–8.46)	0.549	0.79 (0.4–1.55)	0.493	0.58 (0.18–1.86)	0.357
Both	0.24 (0.11–0.54)	0.001	0.67 (0.19–2.33)	0.526	0.62 (0.28–1.38)	0.244	1.05 (0.27–4.00)	0.949	3 09 (0.64–14.82)	0.158	0.00 (0.00–Inf)	0.995	3.34 (1.42–7.87)	0.006	1.63 (0.41–6.51)	0.493
Age of the perpetrator ≤ 30 (Ref ≥ 30)
Yes	2.11 (1.03–4.33)	0.042	2.41 (0.83–7.03)	0.106	0.57 (0.32–1.04)	0.069	0.51 (0.13–1.99)	0.335	0.88 (0.33–2.34)	0.805	7 89 (0.88 70 92)	0.065	0.45 (0.19–1.07)	0.070	0.31 (0.09–1.12)	0.075
Group perpetration (Ref = no)
Yes	0.84 (0.55–1.29)	0.422	1.25 (0.62–2.52)	0.538	0.71 (0.48–1.05)	0.086	0.49 (0.23–1.06)	0.07	1.57 (0.82–2.97)	0.171	0.60 (0.17–2.12)	0.425	2.13 (1.27–3.58)	0.004	3.1 (1.25–7.67)	0.014
Status of the perpetrator (Ref = asylum seeker)
National citizen	0.14 (0.08–0.25)	<0.001	0.28 (0.12–0.67)	0.004	0.95 (0.58–1.56)	0.844	1.06 (0.43–2.62)	0.901	1.29 (0.6–2.78)	0.519	0.69 (0.17–2.75)	0.595	25.91 (13.41–50.07)	<0.001	27.5 (7.58–99.71)	<0.001
Refugee	0.71 (0.32–1.56)	0.390	2.84 (0.72–11.2)	0.135	0.75 (0.37–1.5)	0.414	0.61 (0.18–2.03)	0.416	1.31 (0.42–4.14)	0.641	0.44 (0.04–4.34)	0.483	0.32 (0.04–2.58)	0.283	0.00 (0.00–Inf)	0.996
Undocumented migrant	2.39 (0.53–10.8)	0.258	1.15 (0.30–4.41)	0.84	0.38 (0.17–0.84)	0.017	0.53 (0.13–2.18)	0.379	0.92 (0.20–4.26)	0.919	1.56 (0.26–9.39)	0.626	1.39 (0.29–6.62)	0.679	0.00 (0.00–Inf)	0.997
Sex of the victim (Ref = female)
Male	2.57 (1.65–4.00)	<0.001	0.99 (0.47–2.11)	0.983	0.79 (0.52–1.18)	0.252	1.47 (0.64–3.39)	0.365	0.14 (0.06–0.30)	<0.001	0.29 (0.09–1.00)	0.05	0.93 (0.54–1.59)	0.778	2 (0.73–5.46)	0.177
Both	0.66 (0.34–1.29)	0.228	0.56 (0.21–1.45)	0.229	1.20 (0.59–2.44)	0.618	1.25 (0.43–3.61)	0.677	0.53 (0.19–1.45)	0.216	0.16 (0.02–1.36)	0.094	2.32 (1.08–4.95)	0.030	2.07 (0.61–7.00)	0.24
Age of the victim was ≤ 30 (Ref ≥ 30)
Yes	1.42 (0.64–3.16)	0.389	0.73 (0.18–3.02)	0.666	0.97 (0.42–2.21)	0.935	0.94 (0.21–4.26)	0.934	1.88 (0.39–8.99)	0.429	2.13 (0.17 26 03)	0.553	0.62 (0.24–1.60)	0.327	181293702.45 (0.00–Inf)	0.998
Group victimization (Ref = no)
Yes	0.80 (0.52–1.24)	0.319	1.37 (0.69–2.70)	0.365	1.19 (0.79–1.78)	0.397	0.96 (0.44–2.07)	0.912	1.28 (0.66.–2.47)	0.463	0.75 (0.24–2.35)	0.617	1.84 (1.10–3.06)	0.020	2.28 (0.94–5.53)	0.067
Status of the victim (Ref = asylum seeker)
National citizen	0.97 (0.5–1.88)	0.924	1.51 (0.72–3.18)	0.274	2.01 (0.98–4.12)	0.055	1.68 (0.73–3.86)	0.219	0.59 (0.17–2.04)	0.406	3 04 (0.74–12.56)	0.125	0.10 (0.01–0.73)	0.023	0.04 (0.01–0.21)	<0.001
Refugee	0.69 (0.34–1.40)	0.305	0.47 (0.14–1.53)	0.21	1.02 (0.51–2.06)	0.958	1.3 (0.37–4.62)	0.683	2.58 (1.08 6.15)	0.032	4 72 (0.81 27 67)	0.085	0.39 (0.11–1.32)	0.130	0.52 (0.13–2.04)	0.348
Undocumented migrant	1.73 (0.71–4.20)	0.224	0.94 (0.21–4.09)	0.929	1.39 (0.7–2.76)	0.353	0.6 (0.13–2.92)	0.531	0.21 (0.03 1.68)	0.142	2.99 (0.25 35 46)	0.385	1.57 (0.71–3.50)	0.265	0.00 (0.00–Inf)	0.995

Significant P<0.05 bolded.

Gender

When males commit violence, it is more likely that they engage in sexual violence compared with females, who are more inclined to perpetrate emotional violence. When sexes perpetrate together, it more presumably involves socio-economic and less presumably physical violence than when they act alone or with someone of their own sex. In contrast to the other sex, males are most likely to endure physical victimization, whereas females more probably experience sexual victimization.

Legal status

When asylum seekers commit violence, they more likely engage in physical perpetration than national citizens (here = professionals). They are also more inclined to perpetrate emotional violence compared with undocumented migrants. In contrast to asylum seekers, when national citizens perpetrate, it more presumably involves socio-economic violence. This is echoed in the data on perpetrators rather being a professional than a resident and the victim rather being a resident than a professional (peer aOR: 33.8 [16.54; 69.07], P < 0.001; self-reported aOR 32.77 [9.34; 115.03], P < 0.001). When refugees are victimized, the chances are that they will be sexually victimized compared with asylum seekers. When asylum seekers are victimized, it more likely concerns socio-economical violence in contrast to national citizens. When national citizens are victimized, it more plausibly concerns emotional victimization compared with asylum seekers.

Operation modus

When a group of perpetrators commit violence and when this group consists of both sexes, it more probably involves socio-economic violence compared with violence committed alone or by one single sex. When one is victimized in a group, regardless of its gender composition, it again more likely concerns socio-economic violence compared with being victimized alone.

Reported responses to the reported SGBV cases

In the majority of the violence incidents, someone reacted. Table 4 shows who reacted and what the reactions entailed per type of violence.

Table 4

Reaction to the reported violence cases

Did someone react to the violence incidents?	Yes	Co-residents	Residents and staff	Soc. worker/care worker	Security police/army	Management	Staff management/ security police
N = 1036	832 (80.31%)	116 (11.20%)	157 (15.15%)	158 (15.25%)	114 (11.00%)	131 (12.64%)	156 (15.06%)
Physical violence (N = 437)	363 (83.07%)	47 (10.76%)	68 (15.56%)	73 (16.70%)	54 (12.36%)	53 (12.13%)	68 (15.56%)
Emotional violence (N = 420)	347 (82.62%)	36 (8.57%)	74 (17.62%)	69 (16.43%)	42 (10.00%)	58 (13.81%)	68 (16.19%)
Sexual violence (N = 62)	48 (77.42%)	11 (17.74%)	9 (14.52%)	9 (14.52%)	4 (6.54%)	8 (12.90%)	7 (11.29%)
Socio-economic violence (N = 117)	7 (5.98%)	22 (18.80%)	6 (5.13%)	7 (5.98%)	14 (11.97%)	12 (10.26%)	13 (11.11%)
Reaction consisted of:	Number of reaction	Discussing arranged friends	Interrupting fight calming down	Reporting and investigation	Informing security police army	Arrest	Transfer perpetrator
Physical violence (N = 437)	74 (16.93%)	96 (21.97%)	105 (24.03%)	42 (9.61%)	64 (14.65%)	20 (4.58%)	23 (5.26%)
Emotional violence (N = 420)	73 (17.38%)	101 (24.05%)	84 (20.00%)	39 (9.29%)	69 (16.43%)	15 (3.57%)	29 (6.90%)
Sexual violence (N = 62)	14 (22.58%)	8 (12.90%)	11 (17.74%)	10 (16.13%)	6 (9.68%)	0 (0 00%)	9 (14.52%)
Socio-economic violence (N = 117)	42 (35.90%)	14 (11.97%)	3 (2.56%)	21 (17.95%)	20 (17.09%)	4 (3.42%)	12 (10.26%)

Discussion

Our results confirm earlier literature on vulnerability to SGBV of people with restricted residence permits as asylum seekers, refugees and undocumented migrants,^14–16 as well as people living in detention.^17–20 Yet our results also suggest that living or working in an asylum reception facility is to be considered a risk factor as such. In terms of Desirable Prevention actions, these findings imply that it is paramount to invest in integral prevention actions at the organizational level of the reception settings within the whole European asylum reception sector. While mainstreaming for sexual violence perpetration, it is however advisable to pay attention to preventing asylum seekers from physical and emotional perpetration and to preventing professionals from committing socio-economic violence. Overall, all potential staff members should be better screened on attitudes towards conflict and violence, human rights and discrimination, power indifferences and their coping skills and intercultural competence. Once employed, they need regular training and a code of conduct as part of a violence prevention and response policy that addresses the root causes and triggers of violence rather than consisting of repressive measures. Given the clear group character that socio-economic violence has in this sector, it is vital to address group dynamics in perpetration and build on community resilience when addressing victims. Given the many reports of violence committed by security staff and service providers employed by others, and the fact that residents are regularly transferred from one setting to another as a ‘solution’ to an incident of violence, it is crucial that these policies are imbedded in a sector-wide approach with high-level participation of both residents and professionals.

Another important finding is that victimization and perpetration in this sector seem more gender-balanced than what is generally expected in people outside this sector.^26,35,36 Our results demonstrate that both sexes here perpetrate and experience all types of violence. Moreover, whereas both sexes have a comparable tendency to physical perpetration, a dynamic of mixed-sex perpetration and victimization is to be noted in socio-economic violence. Yet, when they commit sexual and emotional violence, males are more likely to involve in sexual perpetration and emotional victimization, whereas females are more likely to perpetrate emotional violence and experience sexual victimization. This questions the prevailing paradigm in current violence research, in which men are considered a priori the perpetrators and women the victims. Recent research on autochthon intimate partner and domestic violence already pointed to gender dynamics similar to our findings;^37–39 yet, in migration research, this hypothesis has not yet been reflected. This is problematic, because it ignores a number of victims and perpetrators who are in need of effective interventions and who are now left unaddressed. This ignorance leads to ill health consequences and enhances the risk of subsequent perpetration and victimization in current and future generations.^{10,11,21–23} It is thus paramount that future research on violence stems from a gender-sensitive paradigm and reveals all sex and gender dynamics. Consequently, violence prevention actions in the European asylum reception sector should thus avoid messages in which men are stereotyped as sole perpetrators and women as sole victims. Bearing our conceptual framework in mind, we however recommend that these actions are culturally competent, developed and implemented with high-level participation of all types of professionals and residents.

Finally, for all types of violence but socio-economic, only a minority of the significant findings in peer reported victimization were statistically confirmed in the personal victimization cases. This limitation may be due to the possibility of violence being witnessed in asylum centres, increasing the odds of reporting to us by a peer. However, this could also be influenced by residents who feared impact on their asylum case or stay at the facility, as many indicated before, during or after the interview. In addition, in some big facilities and/or communities with honour rules, residents discouraged others to participate, warning of potential stigma and/or community repercussions. Also, professionals indicated not daring to speak openly, although the management had consented to it. In The Netherlands, respondents were recruited and interviewed through the external health care facilities of the reception sector, and feared less disclosure to peers in the centres. This might explain why in The Netherlands, more people disclosed personal victimization in comparison with the other countries. However, disclosing proper involvement was not necessary, and although respondents could disclose indirectly, as is recommended in detention research,²⁰ they were not obliged to respond. This suggests that our findings give a good indication, yet probably still underestimate the real magnitude of violence occurring in this sector. Furthermore, it also indicates that for our research population, trust is a non-evident matter that hampers them from disclosing personally. Therefore, it is important that in future comparable research, respondents are granted the opportunity to both personal and peer reporting, as considering only one of them will only result in revealing parts of the picture. Ideally, retrospective research could complement the findings. The low reporting of violence in Portugal is presumably due to the small number of reception facilities, while our research specifically inquired on violence within those facilities. Another limitation could be the epistemology of the CR, which might have deferred, despite their standard training. Given our findings, and the lack of standard violence prevention policies in this sector, we consider it paramount that professionals, residents and the European citizens proceed to action. As the Senperforto Frame of Reference is freely available in many languages and endorsed by UNHCR, we suggest that when not implemented directly, it is at least consulted as inspiration. It would be interesting to further research the impact of specific housing aspects of the asylum reception facilities on violence occurrence. Finally, understanding how social capital and definitions of violence affect violence reporting would surely build to a better understanding of violence occurrence and its Desirable Prevention.

Funding

European Daphne Fund (JLS/2007/DAP-1/084).

Conflicts of interest: None declared.

Key points

• Living and/or working in the European asylum and reception sector exposes one to violence.

• Both residents and professionals are at risk of both victimization and perpetration, yet they differ in types of violence, targeted victims and perpetration modus.

• Both females and males are at risk of both victimization and perpetration of all types of violence, yet specific characteristics in perpetration and victimization are found.

• When violence occurs in group, it most likely involves socio-economic violence committed by professionals of both genders targeting a group of residents.

• There is an urgent need for mainstreamed, gender-sensitive and culturally competent violence prevention and response actions that stem from a Desirable Prevention approach addressing determinants at the individual, interpersonal, organizational and societal level.

Acknowledgements

Preliminary results were orally presented at the 7th European Congress on Tropical Medicine and International Health, Barcelona 2011, with published abstract; and at the 4th Conference on Migrant and Ethnic Minority Health in Europe, Milan 2012. The Senperforto team consisted of a large consortium, and the authors thank Minor Ndako, JUNA, Fedasil, Rode Kruis Vlaanderen, UNHCR Europe, Nivel, UCD Dublin, JRS Malta, GRC Greece, Menedék, IHMT-Lisbon and ICRH-Ghent University. The authors thank and pay their utmost respect to all respondents of the Senperforto study. Furthermore, they express their enormous gratitude to all community researchers for participating collegiately throughout the project, without them this project could not have been conducted. They also thank the organizations, institutions and persons who took part in the CABs and still try to take Senperforto further in their respective asylum facilities, cities, countries and Europe today.