The experience of violence against children in domestic servitude in Haiti: Results from the Violence Against Children Survey, Haiti 2012

Abstract

Background

There have been estimates that over 150,000 Haitian children are living in servitude. Child domestic servants who perform unpaid labor are referred to as “restavèks.” Restavèks are often stigmatized, prohibited from attending school, and isolated from family placing them at higher risk for experiencing violence. In the absence of national data on the experiences of restavèks in Haiti, the study objective was to describe the sociodemographic characteristics of restavèks in Haiti and to assess their experiences of violence in childhood.

Methods

The Violence Against Children Survey was a nationally representative, cross-sectional household survey of 13–24 year olds (n = 2916) conducted May–June 2012 in Haiti. A stratified three-stage cluster design was used to sample households and camps containing persons displaced by the 2010 earthquake. Respondents were interviewed to assess lifetime prevalence of physical, emotional, and sexual violence occurring before age 18. Chi-squared tests were used to assess the association between having been a restavèk and experiencing violence in childhood.

Findings

In this study 17.4% of females and 12.2% of males reported having been restavèks before age 18. Restavèks were more likely to have worked in childhood, have never attended school, and to have come from a household that did not have enough money for food in childhood. Females who had been restavèks in childhood had higher odds of reporting childhood physical (OR 2.04 [1.40–2.97]); emotional (OR 2.41 [1.80–3.23]); and sexual violence (OR 1.86 [95% CI 1.34–2.58]) compared to females who had never been restavèks. Similarly, males who had ever been restavèks in childhood had significantly increased odds of emotional violence (OR 3.06 [1.99–4.70]) and sexual violence (OR 1.85 [1.12–3.07]) compared to males who had never been restavèks, but there was no difference in childhood physical violence.

Interpretation

This study demonstrates that child domestic servants in Haiti experience higher rates of childhood violence and have less access to education and financial resources than other Haitian children. These findings highlight the importance of addressing both the lack of human rights law enforcement and the poor economic circumstances that allow the practice of restavèk to continue in Haiti.

1. Introduction

Estimates suggest that between 150,000 and 500,000 Haitian children are living in domestic servitude. Child domestic servants in Haiti are known as “restavèks,” derived from French “rester avec” (to stay with), and the term carries demeaning connotations (Balsari, Lemery, Williams, & Nelson, 2010; Centers for Disease Control and Prevention and the Interuniversity Institute for Research and Development, 2011; Hoffman, 2012; McCalla, 2002; Kennedy, 2012; Pierre, Smucker, & Tardieu, 2009; Sommerfelt, 2002). Restavèk children typically come from impoverished, often rural, families who wish to offer their child more opportunity and upward mobility. As such the child is placed into a higher-income, generally urban, home of either strangers or kin with the expectation that the host family will provide for the child’s basic needs and pay for the child’s education in exchange for unpaid labor (Centers for Disease Control and Prevention and the Interuniversity Institute for Research and Development, 2011; McCalla, 2002; Minnesota Lawyers International Human Rights Committee, 1990; Pierre et al., 2009; Restavèk Freedom, 2011; Sommerfelt, 2002). Despite the good intent of families to provide more life opportunities for their children, the experience of child domestic servants exists on a continuum. In the best-case scenario, children live with extended family, experience no maltreatment, perform light household chores and are able to attend school, but in the worst-case scenario, conditions may be more consistent with child slavery (Sommerfelt, 2015).

For example, some studies have found that restavèks are often unable to attend school (Haydocy, Yotebieng, & Norris, 2015) and may work up to 12 h a day performing all the household cooking, cleaning, washing, errands, and other physically demanding tasks, like collecting water (Centers for Disease Control and Prevention and the Interuniversity Institute for Research and Development, 2011; Hoffman, 2012; Kennedy, 2012; Sommerfelt, 2002; McCalla, 2002; Minnesota Lawyers International Human Rights Committee, 1990; Pierre et al., 2009; Restavèk Freedom, 2011). Compared to children of the host household, most restavèks receive inadequate food and clothing, sleep on the floor, and dine separate from the host family. They are expected to address all other household members, including younger children, in formal terms (Centers for Disease Control and Prevention and the Interuniversity Institute for Research and Development, 2011; Minnesota Lawyers International Human Rights Committee, 1990; Pierre et al., 2009; Hoffman, 2012; Restavèk Freedom, 2011; Sommerfelt, 2002). Even if permitted some of the aforementioned privileges, restavèk children still face significant stigma and maintain a social status lower than other children, often being viewed as “property,” and live in fear of abandonment (Centers for Disease Control and Prevention and the Interuniversity Institute for Research and Development, 2011; Hoffman, 2012; Kolbe & Hutson, 2006; Minnesota Lawyers International Human Rights Committee, 1990; Pierre et al., 2009; Sommerfelt, 2002).

Restavèks additionally face many documented risk factors for child maltreatment including lower socioeconomic status, lack of education, living outside of the biological home, and being separated from and not having a relationship with their biological parents (Brown, Cohen, Johnson, & Salzinger, 1998; Breiding et al., 2011; Euser, Alink, Tharner, van Ijzendoom, & Bakermans-Kranenburg, 2013). As a marginalized sub-population of children, restavèks also face unique circumstances which may further increase their risk for experiencing violence such as social isolation; inferior social status; and potentially even a position of servile dependence. There are multiple reports and testimonies from former restavèk children that they faced neglect, physical, sexual, and emotional abuse as a consequence of simply being a restavèk, and one study on human rights violations in Port Au Prince found that female restavèks faced more sexual violence than other girls (Kolbe & Hutson, 2006). However, there has been little quantitative research on violence against restavèk children as compared to other children in Haiti (Centers for Disease Control and Prevention and the Interuniversity Institute for Research and Development, 2011; McCalla, 2002; Restavèk Freedom, 2011). A recent national study (Haydocy et al., 2015) used the 2012 Haitian DHS to estimate the prevalence of restavèk children and compare their living situation to other Haitian children. The study found restavèks had less access to education and more labor responsibilities than other children, but did not experience more physical violence (Haydocy et al., 2015). However, only one adult from the household was interviewed and asked to self-report on perpetration of physical violence against restavèks, so estimates may not be accurate if physical violence was perpetrated by others in the home or the respondent was not comfortable disclosing perpetration. Furthermore, emotional and sexual violence were not examined in this study. Similarly, the national report on the Violence Against Children Survey in Haiti examined demographic variables associated with having experienced violence and found child domestic servitude to be one of them, but did not examine the magnitude of these associations across all age groups nor in the context of specific sub-types of violence, nor demographic variables associated with child domestic servitude (Centers for Disease Control and Prevention and the Interuniversity Institute for Research and Development, 2011). To further expand our understanding of the experiences of restavèks in Haiti, the purpose of this paper is to describe the sociodemographic characteristics of restavèks in Haiti, assess their experiences of childhood violence, and compare these characteristics and experiences to those of youth who were not restavèks.

2. Methods

2.1. Participants

The Violence Against Children Survey (VACS) is a nationally representative, cross-sectional household survey of male and female 13–24 year olds that was conducted from May–June 2012 in Haiti.

2.2. Inclusion/exclusion criteria

Inclusion criteria for this study were males and females living in selected households in Haiti who were 13–24 years of age at the time of the survey and who spoke Haitian Kreyol. Males and females with mental disabilities who did not have the capacity to understand the questions being asked or those with physical disabilities (e.g. hearing or speech impairment) that prevented the interviewer from administering the surveys were also excluded.

2.3. Study design

The study used a stratified multi-stage cluster design including both standard enumeration areas, known as SDE, as well as internally displaced persons (IDP) residing in camps/tent settlements resulting from the 2010 earthquake.

2.4. Sampling techniques

The VACS sample design was first stratified into IDP camps/tent settlements and non-camp enumeration areas. The sampling frame was provided by the Haitian Institute of Statistics and Information (IHSI) and reflected updated projections from the 2003 census. In preparation for this national survey IHSI implemented a rapid count of households using a random sample of SDEs in 2011 to account for expanding cities and communes affected by the January 2010 earthquake. In addition, the Camp Coordination and Camp Management Cluster provides quarterly census data on the IDP camps as a result of the earthquake. Therefore, the most updated sampling frame available was used for this survey.

Each SDE was randomly assigned to males or females so separate prevalence estimates for violence victimization could be calculated and to protect the confidentiality of respondents and minimize the chance that a perpetrator and victim were interviewed in the same community.

For the non-camp sample a 95% confidence interval of +/−2% was desired, and a design effect of two around an estimated 30% prevalence of childhood sexual violence was assumed. This resulted in a required sample size of 1008 males and 1008 females. The estimated number of households needed to generate this sample size for females and males living outside of camps, taking into account both eligibility and non-response at the individual and household level, was 2752 for females and 3415 for males. For the non-camp sample, a total of 177 SDE proportionally allocated by the geographical sub-divisions of Haiti, were selected with probability proportional to size from a total of 11,967 SDE.

Internally displaced persons (IDP) from the 2010 earthquake living in camps were not included in the sampling frame provided by IHSI, thus in order to provide representative national estimates, a separate camp sample was selected. Camp population estimates were made available by the Camp Coordination and Camp Management (CCCM), a partnership of organizations which assists internally displaced persons and manages camps during complex humanitarian emergencies under the leadership of the International Organization of Migration and the United Nations High Commissioner for Refugees. The camp population was estimated to be 5.6% of the total national population, and given that the non-camp sample was fixed at 1008 male and female interviews, the desired camp sample size was determined to be 60 for males and 60 for females. Adjusting for eligibility and non-response at the household and individual levels it was determined that 204 and 164 households would need to be visited for males and females, respectively. A total of 11 camps were selected with probability proportional to size. All but five SDEs were surveyed, as these SDE were not available due security issues.

2.5. IRB review process and consent

Participation in the survey was voluntary and the fieldwork strictly adhered to the study protocol reviewed and approved by independent ethical boards at the Ministry of Public Health and Population’s National Ethics Committee in Haiti, the Interuniversity Institute for Research and Development’s Internal Review Board, and the U.S. CDC’s Institutional Review Board. The protocol included adherence to the World Health Organization guidelines on ethics and safety in studies on violence against women. The interviewers described the study to the primary caregiver in general terms using a list of topics, including health, safety, and community violence, when seeking permission to speak with the respondent. This modified consenting procedure was designed to protect the participants in the event that the head of household was a perpetrator of violence. Informed oral assent was obtained from the respondent. Adult respondents and emancipated minors provided oral consent.

2.6. Setting

All interviews were conducted in a private setting, in or near the household. Interviews were only conducted if privacy could be assured, meaning no one could overhear the interview. All participants were offered a list of organizations that provided local and regional services as well as a national violence hotline. Participants who experienced violence in the last 12 months, asked for help, or became upset in the interview were offered counseling services as well.

2.7. Survey development, description, and administration

The survey was administered in Haitian Kreyol, consistent with previous national surveys in Haiti, and it employed two components: a short demographic interview with the head of household and a comprehensive interview including questions about childhood violence asked of 13–24 year old female or male respondents. The questionnaire took an average of 60 min to administer. The English version of the survey instrument was translated into Kreyol and back-translated into English. The translation was cross-validated with the language translation team and any subsequent corrections identified by this team were made in Kreyol. During the pilot, further corrections were made as necessary. Respondents were asked questions on the following topics: demographics; relationship with parents; education; marriage and partnership; social network and safety; sexual behavior and practices; money and goods exchanged for sex; substance abuse; gender attitudes; HIV testing; current physical and mental health status; physical, emotional, and sexual violence that occurred before age 18; and utilization and barriers to the use of health and social services for those who have experienced violence. Data were collected via paper and pencil.

2.8. Measures

The association of having ever been a restavèk before age 18 and the experience of physical, emotional, and sexual violence was examined.

2.9. Restavèk

Working as a restavèk in childhood was determined by the questions, “Have you ever worked as a restavèk?” and “How old were you when you began working as a restavèk?”

2.10. Sociodemographic variables

Having ever worked for money was assessed by asking, “Have you ever worked for money or any other payment?” School attendance was assessed by asking respondents, “Have you ever been to school?” and if they had, “Have you completed primary school?” Household economics during childhood were assessed by asking, “[Before you were 18 for respondents aged 18–24] do you think your household has/had enough money for a) food b) clothing, school fees, or medical care c) extra money for things like gifts?”

2.11. Lifetime experience of physical violence

Physical violence was defined by asking respondents if a parent/authority figure in the home and/or a public authority figure had ever punched, kicked, whipped, or beat them with an object; choked, smothered, or tried to drown them; burned or scalded them intentionally; and/or used or threatened to use a weapon against them.

2.12. Lifetime experience of sexual violence

Sexual violence included unwanted sexual touching (“touched you in a sexual way without your permission”), attempted sex (“tried to make you have sex without your permission but did not succeed”), pressured sex (“pressured you in a non-physical way by doing things like telling you lies, making promises about the future they knew were untrue, threatening to end your relationship, threatening to spread rumors about you; repeatedly asking for sex or showing they were unhappy; using their influence or authority over you”), and physically forced sex by any perpetrator type.

2.13. Lifetime experience of emotional violence

Emotional violence as assessed by asking respondents if a parent, caregiver, any adult relative, or another adult household member had ever, “said that you were not loved or did not deserve to be loved; that they wished you had never been born or were dead; ridiculed you or put you down; threatened to abandon you or threatened you that they would force you to leave home.”

2.14. Statistical analysis

Data from paper surveys were 100% double-data entered into a CSPro data program (Census and Survey Processing System. Vers. 6.0. Suitland: Census Bureau, 2014. Computer software). Double-data entry required that two different clerks enter data from every survey into CSPro, and a report was run to identify data entry discrepancies. A data manager would then reconcile any inconsistencies by referring back to the paper questionnaire. Data were weighted to represent the Haitian population. SAS Systems for Windows, version 9.3 (SAS Institute, Cary, NC) was used for data management and analysis to account for the sample weights and error due to the complex sample design.

The weighted national estimates, percentages, odds ratios (ORs) and their 95% confidence intervals (CIs) and p values presented were calculated using SAS PROC SURVEYFREQ to account for the sample weights and complex sample design and to calculate the variances of the estimates. The ORs estimate the odds of a given outcome (e.g., lifetime unwanted touching, lifetime forced sex) among those who were ever a restavèk relative to the reference category, having never having been a restavèk. All numbers and 95% confidence intervals (95% CIs) were calculated before rounding. Differences with p values < 0.05 were considered statistically significant.

3. Results

A total of 2902 households (non-camp and camp) were visited for the female sample, of which 1457 females completed an interview. For the male sample 3495 households (non-camp and camp) were visited, and 1459 males were interviewed. The overall response rates were 85.6% and 82.0% for females and males respectively.

In this study the weighted percentage of 13–24 year old respondents reporting they had ever been a restavèk before age 18 was 17.4% for females (weighted n = 225,989) and 12.2% (weighted n = 159,384) for males. Ages of entry into restavèk during childhood ranged from 3 to 17 years old and the mean age of entry into restavèk was 10.3 years old for females, and 9.0 years old for males (Table 1).

Table 1.

Age at which males and females became restavèks, Violence Against Children Survey, Haiti 2012.

Age at becoming a restavèk	Female restavèks	Male restavèks
3–6 years old	51 (18.6% [11.5–25.6])	34 (21.7% [12.4–31.1])
7–12 years old	159 (52.0% [44.2–59.9])	110 (65.3% [55.0–75.5])
13–17 years old	71 (29.4% [22.4–36.4])	26 (13.0% [6.2–19.9])

Data are absolute number (weighted% [95%CI]).

Respondents who had ever been a restavèk in childhood differed from those who had never been a restavèk on a number of key demographic characteristics (Table 2). Restavèks were more likely to have ever worked for money or any other payment. However, because the question on working for money or any other payment was not asked of a specific time frame, the study could determine whether the respondent was compensated for their work while serving as a restavèk. Restavèks were also more likely to have never completed primary school or higher and more likely to not have had enough money for food in childhood.

Table 2.

Lifetime prevalence of childhood (< 18 years old) physical, emotional, and sexual violence among youth ages 13–24 who were ever a restavèk and those who were never a restavèk, by gender, Violence Against Children Survey, Haiti 2012.

	Female				Male
	Ever a restavèk	Never a restavèk	p value	Odds Ratio (95%CI)	Ever a restavèk	Never a restavèk	p value	Odds Ratio (95%CI)
Never worked for money or any other payment	221 (76.8% [71.1–82.5])	980 (85.9% [83.4–88.3])	ref	ref	82 (39.3% [28.6–50.0])	792 (59.3% [55.0–63.5])	ref	ref
Ever worked for money or any other payment	60 (23.2% [17.5–28.9])	170 (14.2% [11.7–16.6])	0.0017	1.83 (1.25–2.67)	88 (60.7% [50.0–71.4])	472 (40.4% [36.2–44.7])	0.0007	2.26 (1.41–3.63)
Never attended school	19 (3.6% [1.6–5.6])	41 (2.5% [1.4–3.7])	0.0008	3.09 (1.56–6.13)	10a (6.6% [2.0–11.2])	21a (1.8% [0.4–3.2])	0.0006	5.96 (1.91–18.57)
Completed less than primary school	179 (67.2% [60.5–73.9])	428 (33.7% [28.8–38.5])	< 0.0001	4.36 (3.13–6.09)	105 (54.3% [43.0–65.5])	522 (35.1% [30.8–39.4])	0.0004	2.49 (1.49–4.16)
Completed primary school or higher	83 (29.2% [22.5–35.9])	681 (63.8% [58.6–69.0])	ref	ref	55 (39.1% [29.2–49.1])	722 (63.0% [58.4–67.5])	ref	ref
Household did not have enough money for food before age 18	191 (71.4% [65.0–77.8])	603 (51.3% [46.0–56.5])	0.0061	2.56 (1.29–5.07)	128 (77.9% [69.7–86.2])	712 (56.2% [51.9–60.4])	< 0.0001	17.52 (3.27–93.72)
Household had enough money for food only before age 18	20a (4.8% [1.6–8.1])	75 (5.8% [4.0–7.6])	0.4340	1.53 (0.52–4.52)	15 (7.2% [3.7–10.7])	102 (8.3% [6.0–10.4])	0.0014	11.02 (1.90–64.05)
Household had enough money for food, medical care, school, and clothing before age 18	33 (11.9% [6.9–16.8])	241 (21.6% [17.5–25.7])	0.9848	1.01 (0.43–2.36)	17a (10.4% [4.0–16.8])	182 (13.5% [11.2–15.9])	0.0031	9.70 (1.59–59.06)
Household had enough money for all of the above and extras before age 18	23 (7.4% [3.9–10.9])	146 (13.6% [10.3–17.0])	ref	ref	2a (0.9% [0.0–2.2])	131 (10.8% [7.7–14.0])	ref	ref

Data are absolute number (weighted% [95%CI]). Absolute numbers do not perfectly correspond to percentages because percentages are weighted.

^aThe relative standard error was more than 30%, the value of estimate may not be stable.

Respondents who had ever served as restavèks in childhood had higher prevalence of nearly every type of violence before age 18 (physical, emotional, sexual) than respondents who had never been restavèks (Table 3). Female restavèks had significantly higher prevalence of childhood physical violence by household members (OR 1.79, [95% CI 1.25–2.57], but this was not the case for males (OR 1.48, [95% CI 0.94–2.31]). Restavèks did not differ from those who were not restavèks on history of childhood physical violence by a public authority figure such as teachers, police, or religious leaders (Female OR 1.34, [95% CI 0.87–2.06]; Male OR 0.78, [95% CI 0.37–1.64]).

Table 3.

Lifetime prevalence of childhood (< 18 years old) physical, emotional, and sexual violence as reported by 13–24 year old respondents who were ever a restavèk versus never a restavèk, by gender, Violence Against Children Survey, Haiti 2012.

Type of childhood violence experienced before age 18	Female				Male
	Ever a restavèk	Never a restavèk	p value	Odds Ratio (95%CI)	Ever a restavèk	Never a restavèk	p value	Odds Ratio (95%CI)
Any physical violence	204 (76.8% [70.4–83.2])	701 (61.9% [56.5–67.3])	< 0.0001	2.04 (1.40–2.97)	118 (69.6% [60.2–79.0])	813 (62.6% [57.5–67.6])	0.1661	1.37 (0.88–2.14)
Physical violence by a household member	191 (70.9% [64.2–77.6])	658 (57.7% (52.9–62.5])	0.0006	1.79 (1.25–2.57)	116 (68.8 [59.3–78.3])	776 (59.9% [54.8–65.0])	0.0894	1.48 (0.94–2.31)
Physical violence by a public authority figure	64 (26.6% [18.5–34.7])	235 (21.3% [15.9–26.7])	0.1662	1.34 (0.87–2.06)	32a (19.4% [8.0–30.8])	310 (23.6% (18.6–28.6)]	0.4972	0.78 (0.37–1.64)
Emotional violence	147 (54.8% [47.2–62.3])	380 (33.4% [29.9–36.9])	< 0.0001	2.41 (1.80–3.23)	89 (51.4% [41.5–61.4])	335 (25.7% [22.2–29.2])	< 0.0001	3.06 (1.99–4.70)
Any sexual violence	116 (39.9% [33.2–46.5])	294 (26.3% [23.3–29.4])	0.0001	1.86 (1.34–2.58)	48 (32.4% [22.6–42.1])	281 (20.5% (17.4–23.6])	0.0111	1.85 (1.12–3.07)
Unwanted touching	79 (26.3% [19.0–33.5])	211 (18.8% [15.2–22.1])	0.0631	1.54 (0.96–2.45)	34 (25.6% [16.3–34.9])	173 (13.5% [10.8–16.1])	0.0037	2.21 (1.25–3.92)
Attempted sex	71 (27.2% [19.9–34.5])	155 (13.0% [10.9–15.2])	< 0.0001	2.49 (1.66–3.76)	18a (13.0% [5.3–20.7])	140 (10.8% [8.5–13.1])	0.5862	1.24 (0.57–2.69)
Pressured sex	25 (8.3% [4.9–11.7])	50 (4.8% [3.0–6.5])	0.0295	1.82 (1.05–3.16)	14a (7.1% [2.5–11.7])	96 (6.6% [5.0–8.3])	0.8527	1.07 (0.50–2.29)
Forced sex	26 (6.0% [2.8–9.2])	62 (6.4% [4.2–8.6])	0.8531	0.94 (0.50–1.79)	10a (5.5% [1.6–9.4])	46 (2.8% [1.8–3.8])	0.0846	2.01 (0.88–4.55)

Data are absolute number (weighted% [95%CI]). Absolute numbers do not perfectly correspond to percentages because percentages are weighted.

^aThe relative standard error was more than 30%, the value of estimate may not be stable.

Both female and male restavèks had significantly higher prevalence of emotional violence before age 18 compared to females and males who had never been restavèks (OR 2.41, [95% CI 1.80–3.23]) and (OR 3.06, [95% CI 1.99–4.70]) respectively.

Females who ever served as restavèks had higher prevalence of any form of sexual violence before age 18 (OR 1.86, 95% CI 1.34–2.58) compared to females who had never served as restavèks in childhood. Male restavèks also had significantly higher prevalence of sexual violence overall (OR 1.85, [95% CI 1.12–3.07]). In terms of specific forms of sexual violence, females who had ever been restavèks in childhood had significantly higher prevalence of attempted sex (OR 2.49, [95% CI 1.66–3.76]) and pressured sex (OR 1.82, [95% CI 1.05–3.16]), while males who had ever been restavèks had higher prevalence of reporting unwanted touching (OR 2.21, [95% CI 1.25–3.92]).

4. Discussion

This study describes the sociodemographic characteristics of male and female restavèks in Haiti, and documents the high prevalence of all forms of violence that restavèks experience. Consistent with prior research, this study documents that restavèks are more commonly female and are less likely than other children to have access to education (Pierre et al., 2009; Haydocy et al., 2015). Approximately 70% of female and male restavèks experienced childhood physical violence and approximately half experienced childhood emotional violence. Female restavèks experienced significantly more physical violence by a household member compared to non-restavèks, but there was not a significant difference in violence by a community member. This is consistent with the fact that most restavèks are not permitted to go to school, as the most frequent perpetrators of community physical violence in Haiti are teachers (Centers for Disease Control and Prevention, Interuniversity Institute for Research and Development, Comite de Coordination, 2014).

Over 30% of females and males who had ever been restavèks in childhood experienced some form of sexual violence before the age of 18. Female restavèks were more than twice as likely to experience childhood physical and emotional violence and close to twice as likely to experience sexual violence compared to female youth who had never been restavèks. Male restavèks were at approximately 50% greater likelihood of physical violence and at over three time’s greater likelihood of emotional violence than non-restavèk males.

Violence against children erodes the foundation children need for leading healthy and productive lives. The high levels of violence experienced by restavèks are likely to make these children particularly vulnerable to a range of negative health and social outcomes throughout their lives. Exposure to violence during childhood can influence subsequent vulnerability to a broad number of mental and physical health problems, ranging from anxiety disorders and depression to cardiovascular disease and diabetes (Centers for Disease Control and Prevention, Interuniversity Institute for Research and Development, Comite de Coordination, 2014; Felitti et al., 1998; Kendall-Tackett, 2003; Shonkoff & Phillips, 2000). It can damage the emotional, cognitive, and physical development of children and, thereby, impact economic development by degrading the contribution of affected children to the human capital of their community and country.

Interventions to protect restavèks are available in Haiti, typically in the form of institutions that support children who were formerly restavèks with shelter, food, and vocational training or education, but little exists in terms of primary prevention of children entering restavèk status (McCalla, 2002; Sommerfelt, 2015). Restavèks are a difficult population to reach, namely because Haitian society perceives the practice as acceptable and often families (both those that provide and receive restavèk children) feel they are actually protecting the child’s best interest by ensuring access to shelter, food, and clothing—resources they may not have otherwise had access to in their family of origin (McCalla, 2002; Blagbrough, 2008; Sommerfelt, 2015). In this regard, cultural rejection of the practice of restavèk is unlikely, which will hinder primary and secondary prevention, without first improving the economic circumstances which make the practice of restavèk the most viable option for a family to ensure their child’s basic needs are met.

Evidenced based intervention strategies exist to prevent violence against children that may have direct relevance to also protecting restavèk children specifically. For example, INSPIRE: Seven Strategies for Ending Violence Against Children references the use of income and economic strengthening as a strategy to prevent violence against children (World Health Organization, 2016). In the case of restavèks, a solution may be to coordinate and concentrate efforts to increase the economic viability of communities identified as being the largest suppliers of restavèk children (Pierre et al., 2009) with programs targeting community health, education, vocational skills, sustainable agriculture, water and sanitation, microcredit, and family planning (Sommerfelt, 2015). UNICEF, for example, addresses these components in its overall human rights based strategy to reduce poverty (UNICEF, 2016). Reducing poverty in communities, combined with social mobilization and awareness campaigns, may facilitate the cultural shift required to end the practice of restavèk (Sommerfelt, 2015).

Additionally the INSPIRE technical package references the importance of implementation and enforcement of laws to prevent violence against children (World Health Organization, 2016) and primary prevention of children being placed in domestic servitude could be achieved through enforcement of the protections afforded to all children of Haiti by virtue of the international human rights laws to which Haiti is a party. These international agreements are designed to protect children from violence, abuse, and exploitation; inhumane and degrading treatment; and labor that interferes with education and is harmful to their health and development. These laws include the Convention on the Rights of the Child; the Universal Declaration of Human Rights; the Supplementary Convention on the Abolition of Slavery, the Slave Trade, and Institutions and Practices Similar to Slavery; and the ILO Conventions Concerning the Prohibition and Immediate Action for the Elimination of the Worst Forms of Child Labor and Forced Labor (International Labour Organization, 1999; UN General Assembly, 1989, 1948; UN Commission on Human Rights, 1986).

Haitian officials have strived to adopt legislation and create institutional frameworks to address child protection (USAID, 2017; U.S. Department of State, 2017). For example, since 2011, the Institut du Bien-Être Social et de Recherche (IBESR) has instituted several improvements and changes. These include a) the creation of IBESR offices in each department of the country; b) the creation of a national database of vulnerable children; c) an effort to conduct a national census of orphanages and d) the implementation of a monitoring mechanism. (INURED, 2017). IBERS is credited with leading Haiti in the ratification of The Hague Convention of May 29, 1993 on Protection of Children and Co-operation on intercountry adoption processes, which officially took effect in 2014, and has worked with the Haitian parliament to update the legal framework on children’s rights by adopting international standards (INURED, 2017). There are currently child protection laws that have been adopted, such as the 2001 law prohibiting corporal punishment, a 2003 law prohibiting all forms of abuse against children and a 2005 law establishing mandatory minimum sentencing of 15 years hard labor for child sexual abuse (CRC, 2015). Furthermore, under Haitian law, it is prohibited to employ children under the age 15 and it is mandated that all children have the right to go to school (Sommerfelt, 2015). Progress is being made to address child domestic servitude in Haiti with efforts such as the development of a Child Protection Code and formation of a National Committee to Combat Human Trafficking, but coordination of efforts from the international community and Haitian civil society through a multisectoral action plan is still required (Sommerfelt, 2015).

Despite these efforts, the state has no capillary power to enforce the many child protection laws and regulations it has endorsed. As this traditional practice of child fosterage has evolved—and traditional kinship practices and hierarchies that once served as protective measures have been eroded by poverty, coupled with migration to and from Haiti, as well as the recurrence of disasters—the State has been unable to step into the role once played by kinship networks, subjecting children to exploitation (Marcelin and Cela, 2017).

Prior to this survey, there was no national level data on restavèks, so estimates on the number of restavèks ranged widely from 150,000 to 500,000. The weighted number of restavèks in this study was 385,373, which is about 15% of the population of 13–24 year olds in Haiti in 2012. Yet, due to the nature of the survey design, the estimates for the size of the population of restavèks in Haiti and their experience of violence are likely to be underestimates. One reason is that current restavèks may have been under sampled in the final stage of selection if the head of household (HOH) did not include the restavèks in household listings, for example if the HOH does not perceive the child to be a member of the household. There may also have been differentially higher rates of refusal on the part of the guardian to allow that particular child to complete the survey due to fear of what the restavèk might share or due to lack of willingness to let them stop working for the duration of the survey. Former restavèks may also have been under-sampled if they are less likely to be living in a household versus an institutional setting, boarding house, workhouse, orphanage, or the street. Furthermore, among current or former restavèks who were survey respondents, there may have been a differentially higher rate of withholding information on violence due to fear of retaliation, leading to underestimates of violence.

Limitations of this analysis in regards to the survey itself include a lack of information on current status as a domestic servant in their household, as the survey only asks if the child has ever been a restavèk before age 18 (not if they are currently a restavèk). Due to the sensitive nature of this topic and the derogatory association with domestic servitude in Haiti, the question was worded in this manner to best ensure protection of the child and obtain honest responses. However, this limits the ability to draw inferences to their status of restavèk when individual incidents of violence occurred. For example, it is possible that young victims of violence are at increased risk of becoming a restavèk and that violence occurred before a child became a restavèk. In this analysis, we therefore present only associations and are unable to comment on the temporality of such experiences.

Despite these limitations, this analysis represents the most comprehensive quantitative, nationally representative report on the demographic characteristics of restavèks and their experiences with violence. These findings underline the value of data on this vulnerable population and information that can guide public health actions to address their needs. The nationally-representative survey design afforded a unique opportunity to speak with restavèks in a non-targeted and thus more discreet manner, and provided access to this population on a level geographically and quantitatively not previously accomplished. To complement this approach, further studies specifically targeting child domestic servants in order to oversample this population and further describe the breadth of their experiences are recommended.

This study helps to make visible the sociodemographic profiles and experiences of violence endured by restavèks in Haiti, a population which is often hidden from view and whose human rights are often overlooked due to their social status. This study provides information about the magnitude and nature of violence against restavèks which will be useful for policy-makers, specialists in child protection, and educators, among others, in helping to build new and strengthen existing programs and policies designed to address child domestic servitude in Haiti.