Robin Wilson. Handbook of Children, Culture, and Violence. Editor: Nancy E Dowd, Dorothy G Singer, Robin Fretwell Wilson. Sage Publication. 2008.

Myths surrounding child sexual abuse hinder the ability of judges and others to protect children from real threats they face. One particularly deep-rooted and damaging myth maintains that incest with one child is an isolated event that a parent is not likely to repeat with another child. This chapter argues that the perpetrator who commits incest rarely stops with the first victim. To the contrary, incest is a predictably recurring event that could largely be avoided if child protective services (CPS) agencies would remove the alleged perpetrator from the home, rather than removing the victim while leaving other children in the home.

The choice to remove the child rather than the threat is driven largely by misunderstandings about the legality of excluding alleged offenders from their home, compounded by the equally entrenched, but wrong-headed view that a non-abusing parent who fails to protect once will do so again. CPS legally can, however, and should, place the burden of homelessness on the alleged offender rather than compromising children’s safety.

Clouded Judgment

When a parent sexually abuses a child in his or her care, a question frequently arises regarding the safety of other children in the household. Because child sexual abuse by women occurs very rarely (Wilson, 2002), this chapter will consider only incest at the hands of fathers and father-substitutes. The state may intervene to protect the victim because he or she has already been harmed. For the state to intervene to protect the victim’s sibling, however, the state must show that the sibling “more probably than not” faces substantial risk of imminent harm. Once the state proves that further incest is probable, it may act to protect additional children in a variety of ways, including removing the child, supervising the family, or mandating “voluntary” treatment of the perpetrator (Fla. Stat. Ann. § 39.52(1)(b), 2003; Goldstein, 1999).

Perhaps because judges believe the isolated act myth, they reach wildly different judgments regarding the risk to children left in the perpetrator’s care. Courts in the United States generally react in one of three ways.

No Clear Risk

Some courts see no clear risk to the victim’s siblings. A New York family court in In re Cindy B. (1983) refused to protect the siblings of an incest victim, finding that the state produced no evidence “that the physical… condition of any [sibling] … is in imminent danger of becoming impaired” (p. 195). The father admitted sexual intercourse with his oldest daughter, Cindy. The New York Court of Appeals validated this approach in a case where the 12-year-old victim, Starr, was digitally penetrated by her mother’s live-in boyfriend while he “instructed her to lick his penis ‘like an ice cream cone—” (In re Starr H., 1989, p. 767). The state CPS agency petitioned to protect Starr and her siblings, but the family court dismissed the petition without explanation. On appeal, the court found that while Starr was an abused child, her sexual abuse—standing alone—was insufficient to find that her siblings were at substantial risk. New York courts are not alone in refusing to protect a victim’s siblings. Texas courts have reached similar conclusions in family court proceedings to terminate parental rights (Lane v. Jefferson County Child Welfare Unit, 1978). Rather than perceiving a threat to other children, these courts see sex with one child as an isolated instance—a fluke—as opposed to critical evidence of a larger and foreseeable pattern of prédation.

Obvious Risk to the Victim’s Siblings

Other courts treat the risk to siblings as self-evident. The Ohio Court of Appeals in In re Burchfield (1988) held that “a child should not have to endure the inevitable to its great detriment and harm in order to give the [parent] an opportunity to prove [his] suitability” (p. 333). The father had inserted his finger into the vagina of his five-year-old daughter on two separate occasions and the court concluded “in light of [the daughter’s sexual abuse], it follows that so long as the father was in the home with [her siblings] the environment of these children was such as to warrant the state to assume guardianship” (p. 333). As the court explained, “The law does not require the court to experiment with the child’s welfare to see if he will suffer great detriment or harm” (p. 333). A number of other courts also see this risk as a “no-brainer,” including courts in Arizona, California, Oregon, Rhode Island, Nebraska, and South Dakota (In re Appeal in Pima County Juvenile Dependency Action No. 118537, 1994; In re Daniel B., 1994; In re Dorothy I., 1984; In re J.A.H, 1993; In re M.B, 1992; State ex rel. Juvenile Department v. Smith, 1993).

Prior Victimization is Merely a Factor in Deciding Risk to Siblings

A third set of courts view the victim’s violation as a relevant, but not sufficient, factor in determining whether siblings face a substantial risk of harm. The Florida Supreme Court adopted this approach where a father had intercourse with his stepdaughter who was under 12 at the time (In re M.F., 2000, p. 1191). Following his incarceration, the state CPS agency sought to remove the father’s two biological children from their mother’s care based in part on the possibility of future abuse by the father. In a sharply divided opinion, the Florida Supreme Court announced that a parent’s commission of a sex act with one child was, by itself, insufficient to support a ruling of dependency as to the victim’s siblings.

Even when judges agree about the risk to the victim’s siblings, they often sharply differ regarding whether a sibling’s gender, age, ordinal position, and genetic relatedness mute the risk to him or her (Wilson, 2002). For example, courts disagree about whether a father who molests a daughter poses a threat to his sons. Some courts have concluded that a father who molests a daughter will confine his attentions to other daughters and poses no threat to his sons, while others reach exactly the opposite result (In re Burchfield, 1988; In re Rubisella E., 2000). Courts also clash over whether a male parent who molests his stepchild is equally likely to victimize his own flesh and blood (In re S.G., 1990; State ex. rel. Juvenile Department v. Rhoades, 1985).

Empirical Data Supporting Removal of the Threat

Despite these conflicting views on the risk of future offenses, all courts can call on considerable social science data regarding incest to better protect the victim and the victim’s siblings.

Unmistakable Evidence of Risk to Siblings

The evidence of serial offending is overwhelming and chilling. Herman and Hirschman (1981) conducted a study of 40 families containing allegations of father-daughter incest. Victims in 53% of the families reported another victim or “strongly suspected” that incest with a sibling also occurred (p. 94). Forty-seven percent said there was no indication of other victims. However, in one-third of the families studied, there were no other possible female victims in the household. Similarly, in Russell’s (1986) landmark study of 930 women in San Francisco, she found that one-half of the children abused by a stepfather reported at least one other victimized sibling, while one-third of the women abused by a father reported other sibling-victims. Although alarming, these figures may actually underestimate the incidence of serial prédation due to the intense secrecy surrounding incest and the common assumption by victims that they are alone in being molested (Russell, 1986). Färber, Showers, Johnson, Joseph, and Oshins’ (1984) study of medical records in 162 molestation cases yielded a lower rate of repeat incest with another child, 28%. While 72% of the records examined gave no indication of additional incest, in 41% no one asked the victim whether others may have also fallen prey.

The pattern of multiple victims in the same household is hauntingly familiar. De Francis (1969) studied 250 sexual abuse cases and found that in 22% of the cases, perpetrators victimized between two and five children. Faller (1990) analyzed 196 paternal caretakers whom she classified in two ways: biological father-offenders and father-substitutes, including stepfathers, mother’s cohabitants, and mother’s boyfriends. Faller indicated that four-fifths of the biological fathers abused more than one child in the household, as did two-thirds of father-substitutes. In many cases, every child in the household had experienced incest. Phelan (1986) found similar results in her study of 102 cases of father-daughter incest. Of the offenders, biological fathers molested 85% of all daughters available to them, while stepfathers molested 70%.

In fact, incest perpetrators have been found to frequently assault several children in their care. In a study of 373 incest offenders, Ballard et al. (1990) constructed a profile of perpetrators that included abuse history. They found 33.9% had at least one additional incestuous relationship after the first. Although frightening on the surface, perhaps more terrifying is how this number breaks down. The largest subgroup, 12.8%, had one additional incestuous relationship; but the second-largest category, 8.4%, represented perpetrators who admitted five or more additional incestuous relationships. Not surprisingly, Ballard concluded that incest offenders “often have histories of large numbers of victims” (p. 46).

Although the risk to siblings is clear, not all children face identical risks. The picture of risk is complex and depends on a number of factors, including the gender of the victim and the siblings, as well as the age of onset of the victim’s abuse (Wilson, 2002). Certain children face only a slim chance of becoming victims. Specifically, there is minimal risk following father-daughter incest that began in the daughter’s teenage years, when the family contains only one other child, a son. Absent other indicators of risk, the male child in this household is not likely to be victimized (Wilson, 2002).

Given the numerous studies of serial victimization, it would seem that the risk to siblings would be obvious. Nonetheless, some early studies of recidivism among incest offenders suggested that an offender, once caught, would just stop (Finkelhor, 1986). These studies predicted that only 4 to 10% of incest offenders would be recidivists (Quinsey, Lalumière, Rice, & Harris, 1995). New studies now suggest that incest offenders remain a continuing threat. Before assessing this new research, it is important to review the early studies as they offer important glimpses of sibling risk that have been overlooked.

In the early studies, incest offenders appeared at first to be much less threatening than sexual offenders who struck outside the home. Sturgeon and Taylor’s (1980) study of 260 mentally disturbed sex offenders, for example, compared the reconviction rates of heterosexual pedophiles; homosexual pedophiles; and incestuous offenders, whether heterosexual or homosexual. Reconvictions for sexual crimes were 20% among heterosexual pedophiles, compared to 15% for homosexual pedophiles, and 5% for incest offenders. According to this comparison, incest offenders appeared to present only modest risks of reoffending. Yet, other evidence in the same study undercuts the incest perpetrator’s image of relative safety. Specifically, 19% of incest offenders had prior convictions for sexual crimes. Although prior convictions for incest offenders fell significantly short of heterosexual pedophiles (43%) and homosexual pedophiles (53%), the findings nonetheless confirm that significant numbers of incest perpetrators do indeed engage in a pattern of repeat offenses against children.

Even before new studies emerged showing serial offenses, researchers faulted these early findings. Larson, Terman, Gomby, Quinn, and Behrman (1994) noted that recidivism is “extremely difficult to measure because many sex crimes may not result in arrest or conviction [and because] … official data are often inaccurate or outdated” (p. 10). Recidivism studies yield misleading appraisals of risk because they typically follow incarcerated offenders. Yet we know that incest offenders generally are not incarcerated (Bolen, 2003; Finkelhor, 1986). Finally, the early studies tracked subjects for short periods and simply missed new offenses occurring many years later, which frequently occur with child molesters (Meyer & Romero, 1980).

Recent studies directly take issue with the old thinking that incest offenders will not reoffend. Studer, Clelland, Aylwin, Reddon, and Monro (2000) grouped 220 patients who participated in an Alberta, Canada, treatment program for sex offenders into those whose index victim was related (incestuous offenders), and those who had abused an unrelated child (extrafamilial abusers). The authors compared the rates at which each reported offending occurred against other children within and outside the home. Contrary to conventional wisdom, they found “22% of the incestuous group had prior offenses against a related child,” suggesting that “repeat offenses may not be so rare” (p. 18). By comparison, only 12.9% of offenders who victimized an unrelated child reported violations against related children, making incest offenders nearly twice as likely to report other related victims. As Studer et al. note:

[I]f the “dogma” [of the incest offender’s low propensity to reoffend] were theoretically and clinically sound (incest offenders being an entirely separate and discrete group), the [reported rate of other related victims among incest offenders] should approach 0%…. The fact that [0%] is so far from [the reported value] says as much as any real differences [between incest offenders and non-incestuous ones]. (Wilson, 2002, p. 261)

Studer et al.’s findings of continuing risk are mirrored in a raft of recent studies attacking the early distinction between incest offenders and other child molesters. These studies have found that incest offenders and child molesters who strike outside the family have “very similar arousal patterns” (Barsetti, Earls, Lalumière, & Bélanger, 1998, p. 283), indistinguishable erotic preferences (Studer, Ay 1 win, Clelland, Reddon, & Frenzel, 2002), and “disturbingly high” deviant sexual arousal from children (Firestone et al, 1999, pp. 512-513). Many child abuse researchers now question the extent to which “different categories of offenders, particularly intrafamilial and extrafamilial, are different from each other” (Salter, 1988, p. 49) and argue that the classification of sex offenders into two groups, incest offenders and pedophiles, was “prematurely disseminated as [it does] not appear to be valid” (Conte, 1999, p. 25). In the face of this evidence, the antiquated view that incest offenders are a special category who will not reoffend must be discarded.

While it is true that some fathers who victimize one child will not go on to victimize another in the household, the state may act to protect additional children if they face a substantial risk of imminent harm. The studies presented here suggest that the state can readily make this showing. More fundamentally, however, if the father’s act of abuse with the first child is substantiated, the consequences of that abuse—homelessness— should fall on the father rather than the victim and other children, as this chapter argues more fully below.

Distrust of the Non-Abusing Parent

Another factor that drives the policy of removing children rather than the threat is the belief that the non-abusing parent is complicit in the abuse. Removing the children would make sense in this instance because leaving them with a parent who failed to protect the victim might result in more harm. The belief that mothers are complicit is widespread among social workers, yet the available empirical data fail to support that belief. In addition, there are mechanisms to evaluate the non-abusing parent’s ability to protect that are far superior to the solution of removing the child rather than the threat.

Numerous studies show that most caseworkers fiercely believe mothers share blame for incest. A series of studies in the 1990s found that 70 to 86% of all CPS professionals placed some responsibility on mothers, both for incest and for extrafamilial sexual abuse (Johnson, Owens, Dewery, & Eisenberg, 1990; Kelley, 1990; Reidy & Hochstadt, 1993). Some studies asked caseworkers to assign relative responsibility for the abuse. In several of these, the mothers’ perceived responsibility for the abuse ranged from 11% to 21% (Kalichman, Craig, & Follmgstad, 1990; Kelly, 1990). In Australia, Breckenridge and Bal dry (1997) found that 61% of CPS workers felt that some mothers knew of the abuse. One in ten believed that most mothers actually knew about the abuse. In the United States, Ryan, Warren, and Weincek (1991) found that in 82.3% of the case reports from five state, county, and private welfare agencies, caseworkers believed the mothers knew about the abuse before it was reported.

These suppositions of “maternal culpability” (Bolen, 2001, p. 193) translate directly into the choice to remove the child. In Ryan et al.’s (1991) study, assessments of “mother’s ability and willingness to protect her child (1) before and (2) after the report of abuse … best explained] the pattern of removal” (p. 132).

Yet, there is little support for this belief. As Ryan et al. (1991) flatly observes, “Although the myth has been widely held that [the non-abusing mother] is usually aware of the abuse and may contrive in setting it up, this is infrequently the case” (p. 124). In a study of 65 cases of paternal incest, Faller (1990) found that a mere 5% of mothers knew about the daughter’s abuse, but “felt powerless to stop it” (p. 67). A study of grandfather incest found that 87% of mothers never knew. (Margolin, 1992). In 1985, Myer (1985) found that 75% or more of mothers did not know of their partner’s abuse.

Many child victims never tell anyone of the abuse, especially when incest is involved. Mian, Wehrspann, Klajner-Diamond, LeBaron, and Winder (1986) found that the rate of purposeful disclosure by children decreased significantly when the perpetrator was intrafamilial. In fact, a greater proportion of children victimized by family never tell anyone of the abuse (17.7%), as compared to children who are the victims of extrafamilial abuse (10.9%) (Fischer & McDonald, 1998).

While a child’s disclosure may not be not the only clue that abuse is occurring, other cues one might expect are also frequently absent. A third of sexually abused children have no apparent symptoms (Kendall-Tackett, Williams, & Finkelhor, 1993). Roughly half fail to display the classic, most characteristic symptom of child sexual abuse: “sexualized” behavior (p. 167). As disquieting as it is, “the more severe cases [are] the ones most likely to remain secret” (Russell, 1986, p. 373). Russell reports that in 72% of the cases in which mothers were unaware of the abuse, more severe abuse had occurred. All of this makes one wonder how precisely mothers could have ferreted out their children’s abuse. Clearly, “[m] others cannot report what they do not know” (Bolen, 2001, p. 190). This is not to say that mothers can never be complicit in a child’s abuse. They can. Nonetheless, absent specific indications of a mother’s complicity, caseworkers should generally assume that mothers did not simply go along.

Nor is there any reason to believe that non-abusing mothers are not protective after the abuse. Most are “very” or “mostly” protective once they find out. Ninety-one percent of non-abusing mothers in a 2005 study were “supportive” following the disclosure of child sexual abuse (Alexander et al., 2005). Pellegrin and Wagner (1990) found that 74% of non-abusing mothers “either totally or largely believed the child’s account of abuse” (p. 57), while caseworkers rated 67% of mothers as having average or better compliance with the caseworker’s recommended treatment plan. Even Ryan et al.’s 1991 study, in which caseworkers harshly assessed mothers’ knowledge, found that over half the mothers (50.8%) acted “mostly” or “very” protective following the report. Importantly, most mothers believed the disclosure. Sirles and Franke (1989) discovered that 78 % believed the child’s report of alleged abuse. Although there are studies showing that only one in four non-offending mothers were “very supportive” (Adams-Tucker, 1982, p. 1252), such studies are in a distinct minority (Bolen, 2002). One meta-analysis concluded that “75% of nonoffending guardians are partially or fully supportive after disclosure” (Bolen, 2002, p. 40).

In any event, if an unspoken concern that a “mother who failed once will fail again” is informing the decision to remove children, caseworkers should assess the likelihood of a failure prospectively, on the basis of validated assessment tools, rather than the fact of the child’s past abuse. Such tools exist (Bolen, 2002, p. 40) and are widely used elsewhere and in some U.S. jurisdictions (New Zealand Child, Youth and Family Services, 2004; M. Testa, personal communication, Oct. 29, 2004). Caseworkers should also realize that if the alleged abuser is gone, it lessens significantly the burden on the mother, making the question of whether she will do the right thing less of an issue.

Constructing a Safer Path

These studies alone justify a presumption that a perpetrator who strikes once within the family will strike again. The studies affirm, moreover, that most non-offending mothers did not know and could not have known about the abuse before it was disclosed, and that they act protectively of their children following disclosure.

In addition to this evidence, there are a number of sound public policy reasons for presuming risk to other children in the family. First, a presumption of risk sanctions the efforts of CPS caseworkers who, without clear guidance, may be slow to react or may not act at all. In addition, a presumption places the burden on the offender to prove the sibling’s safety, erring on the side of additional protection for other children. After all, the offender chose to sexualize his relationship with one child in his care and he is the primary determinant of repeat performances. Finally, presuming risk gives courts discretion to reject the rebuttal if they sense risk to the siblings rather than requiring additional harm before acting.

Beyond shifting burdens of proof, or improving judicial predictions of risk outlined elsewhere (Wilson, 2002), we should embrace fundamental change: The alleged offender should be removed from the home, pending a full investigation, rather than plucking the victim and other children from their homes. The next section explores why this seemingly radical shift in our default position at the inception of a child abuse investigation is rarely made, but eminently achievable and well within the mainstream of governing legal precedent (Wilson, 2005).

Understanding Decisions to Remove Children

Questions of safety drive the impulse to remove the victim first and sort things out later. Prosecution rates for those who victimize children are abysmal—“93% or more of all offenders are allowed to remain within the child’s environment or to return within the year” (Bolen, 2001, p. 258). Less than 2% of all suspected offenders are convicted, while only 7% of all offenders whose abuse is substantiated ever spend more than a year in jail (Bolen, 2001). This failure to prosecute and convict creates a perceived need to protect children by removing them from their home rather than removing the threat.

In a study of factors influencing the state’s decision to remove a child, Cross, Martell, McDonald, and Ahl (1999) found that “the decision not to prosecute was the strongest predictor of child placement” (p. 41) outside the home. As the authors note, “[i]f cases are not accepted for prosecution … the child’s removal from the home … may be the only way to protect the child” (p. 41). In this instance, child placement is seen “as the lesser of the two evils” (p. 42).

This Hobson’s choice grows out of a deep-seated belief that offenders cannot legally be excluded from their homes absent prosecution—despite the fact that states can, and do, remove children from their homes every day (American Prosecutors Research Institute, 2004, p. 279). Bagley and King (1990) have argued that “[a] proper legal framework which would enable the child to remain with her mother while the alleged offender is removed, still has to be established” (p. 101). Another child abuse researcher, Rebecca Bol en (2003), observes:

Removing the alleged offender instead of the victim from the child’s environment… may be one of the most difficult policy changes because it conflicts with society’s presumption that the accused is innocent until proven guilty, (p. 1358)

The experience of other countries, coupled with the experiences of some U.S. jurisdictions, supports presumptive removal of the alleged offender. Authorities in Great Britain, for example, are explicitly authorized during an investigation to require the accused parent “to leave a dwelling-house in which he is living with the child” (Children Act 1989, ñ 41 § 38A (3)). Indeed, this is the “preferred course of action” (Wickham & West, 2002, p. 153) when a child is at risk from someone living in the home.

Seven American jurisdictions explicitly authorize CPS agencies to obtain protective orders directing an alleged offender to vacate the home, including Hawaii, Kentucky, Maine, New York, Tennessee, and Texas, as well as Guam (19 Guam Code Ann. § 13316, 2004; Haw. Rev. Stat. Ann. § 587-53(f), 1999; Ky. R. Jefferson Fam. Ct. Rule 6 app., 2003; Me. Rev. Stat. Ann. tit. 19A §§ 4005(1) & 4006(5), 1998; Me. Rev. Stat. Ann. tit. 22 § 4036(1)(F-1), 1998; Morgan & Gaither, 1999; N.Y. Fam. Ct. Act § 842, 2004; Tenn. Code Ann. § 37-1-152, 2001).

Statutes Authorizing Removal of the Alleged Offender

These new statutes do not rely on a household member (like the child or mother) to ask for assistance; instead they permit judges and caseworkers unilaterally to remove the offender. Thus, for instance, Maine authorizes its Department of Human Services to petition for a protective order on behalf of a child who has been abused by a family member and allows the court, without the parties present, temporarily to enjoin the alleged abuser from “[e]ntering the family residence” (Me. Rev. Stat. Ann. tit. 19A, SS 4005(1) & 4006(5), 1998). After a hearing, this order may be made permanent for up to 2 years (Me. Rev. Stat. Ann. tit. 19A, SS 4005(1) & 4007, 1998). Tennessee authorizes its CPS agency to apply for a “no contact order” removing the alleged perpetrator from the child’s home if there is probable cause that the adult sexually abused the child (Tenn. Code Ann. § 37-1-152, 2001). Other states also authorize state agencies to take such steps (Haw. Rev. Stat. Ann. § 586-3(b)(2), 1999; Morgan & Gaither, 1999).

In several states and territories, removal of the child can occur only after the court first gives “due consideration to ordering the removal… of the alleged perpetrator from the child’s family home” (19 Guam Code Ann. S 13316, 2004). In both Hawaii and Guam, the “burden of establishing that it is not in the best interests of the child that the alleged perpetrator be removed from the family’s home” (19 Guam Code Ann. § 13316; see also Haw. Rev. Stat. Ann. S 587-53(f), 1999) falls on the child’s family, not the state CPS agency. In Texas, if the state CPS agency determines that “the child would be protected in the child’s home by the removal of the alleged perpetrator,” it “must file a petition” to exclude the alleged offender (Morgan & Gaither, 1999). The court then has no choice but to exclude the parent if it finds that the child has been sexually abused and “there is substantial risk” he or she will be abused again if the parent remains in the residence (Morgan & Gaither, 1999).

These emerging statutes do not simply duplicate the protection already available under domestic violence statutes, although many of the latter would also be available to protect children (Klein & Orloff, 1993, p. 820). Domestic violence statutes generally require that someone says “protect me” (Ky. Rev. Stat. Ann. § 403.725, 2003; Klein & Orloff, 1993; W. Va. Code Ann. §§ 48-27-305 & 48-27-204, 2004). In contrast, these statutes permit CPS agencies unilaterally to act to protect children from the threat in their home.

Domestic violence statutes are also intended “[t]o allow family and household members who are victims of domestic abuse to obtain expeditious and effective protection against further abuse” (Me. Rev. Stat. Ann. tit. 19A, § 4001(2), 1998). Any protective order issued under such a statute is granted for a limited time only (Cooke v. Naylor, 1990, p. 379; Me. Rev. Stat. Ann. tit. 19A, § 4007(2), 1998). In Cooke, the court took an opportunity to explain the difference. It cautioned counsel that protective orders are “not the most efficient use of litigation resources for the final resolution of the controversy” (p. 379) over access to the child. As the court explained, “once a temporary order safeguarded the child from immediate harm” (p. 379), proceedings to assure the child’s safety permanently—as CPS proceedings do—should have followed.

Immunity Doctrines

Even without specific statutory authorization, caseworkers outside these seven jurisdictions also may take steps to remove the alleged offender. A pragmatic barrier, however, might be the risk of liability for social workers from parents if removal is not justified (Pearson, 1998). Courts’ treatment of social workers, however, does not substantiate this risk. To the contrary, case law indicates courts will defer to the decisions of social workers absent egregious circumstances and generally insulate caseworkers from liability when removal is ordered.

One way courts protect social workers is through the application of immunity doctrines. Some courts have accorded absolute immunity similar to that given to judges, to social workers in the performance of certain duties, largely so that they are “free to exercise their discretion without fear of personal consequences” (English, 1995, p. 768). Without such insulation, “[individual caseworkers and supervisors facing the possibility of losing their life savings in a law suit might allow fear to influence their decisions, intentionally or otherwise” (Gottlieb v. Orange County, 1994, p. 629).

Other courts provide a more limited form of immunity, known as qualified immunity. State officials acting under this more limited form of protection still enjoy broad protection from civil liability. Under this framework, a social worker receives qualified immunity when he or she acts, in the words of one court, on the basis of “some reasonable and articulable evidence giving rise to a reasonable suspicion that a child has been abused or is in imminent danger of abuse” (Croft v. Westmoreland County Children and Youth Services, 1997, p. 1126) or, in the words of another, upon “an objectively reasonable suspicion of abuse” (Puricelli v. Houston, 2000, p. 19). If such a basis exists, CPS will be justified in removing either a child or a parent from the home, “even where later investigation proves no abuse occurred” (Croft v. Westmoreland, 1997, p. 1126). The basis for this immunity lies in the balancing of parental and children’s rights:

The due process clause of the Fourteenth Amendment prohibits the government from interfering in familial relationships unless the government adheres to the requirements of procedural and substantive due process: In determining whether [a parent’s] constitutionally protected interests were violated, we must balance the fundamental liberty interests of the family unit with the compelling interests of the state in protecting children from abuse. (Croft v. Westmoreland, 1997, pp. 1125-1126)

The rights of parents with respect to their children, although fundamental, are not without bounds. Instead they are

limited by the compelling governmental interests in the protection of children— particularly where the children need to be protected from their own parents … The right to familial integrity, in other words, does not include a right to remain free from child abuse investigations … Whatever disruption or disintegration of family life [a parent] may have suffered as a result of [a] child abuse investigation does not, in and of itself, constitute a constitutional deprivation. (Croft v. Westmoreland, pp. 1125-1126)

Some courts appear to ratchet up the level of protection further, with an articulated standard that exceeds the “reasonable suspicion” qualified immunity standard:

[A] social worker acting to separate parent and child rarely will have the luxury of proceeding in a deliberate fashion, as prison medical officials can. As a result, in order for liability to attach, a social worker need not have acted with the “purpose to cause harm”, hut the standard of culpability for substantive due process purposes must exceed both negligence and deliberate indifference, and reach a level of gross negligence or arbitrariness that indeed “shocks the conscience. (Miller v. City of Philadelphia, 1999, pp. 375-376, emphasis supplied)

Importantly, in analyzing claims of due process violations by “excluded” parents, courts give equal weight to the interests of parents and children. These courts have not crafted heightened, stringent tests to protect parents from removal.

Application of Immunity Doctrines: CPS Agencies Enjoy Wide Latitude

Of course, tests like these are abstractions. It is their application to specific facts that illustrates just how much latitude courts have given caseworkers. Courts have typically been generous in the application of immunity. For example, in Gottlieb v. Orange County (1994), caseworkers for the county and CPS directed a father to either leave his home because of his alleged abuse of his daughter or face her removal. The father exited for approximately 1 month and later sued, alleging violations of his civil rights. The court began its analysis with the father’s claims against the caseworkers. It found the caseworkers had an objectively reasonable basis for acting, and were therefore immune from suit, even though they made a number of significant missteps: They never investigated the anonymous informant’s background or motives; failed to question the daughter in a neutral, nondirective manner; and asked “neither the daughter’s teacher nor the school nurse, if the child exhibited any behavioral oddities” (Gottlieb v. Orange County, 1994, p. 630). The court refused to fault the workers because they had not been trained in less suggestive means of interviewing.

The father also sued the county and its Department of Social Services (“Department”). While the lower court initially denied their requests for dismissal (Gottlieb v. Orange County, 1994, p. 630), the court ultimately granted the county summary judgment in a later round of litigation, based on undisputed evidence that the county adequately trained its caseworkers (Gottlieb v. Orange County, 1995, p. 73). The Department also acted reasonably, the court found, in issuing an ultimatum to exit without “pausing to obtain a court order” since their source reported ongoing abuse, and since the daughter herself described repeated molestations at her father’s hands, said that her father did not like tattlétales, and said that she expected to be punished for talking about it outside her home (Gottlieb v. Orange County, 1996, p. 520). In the final analysis, the father did not prevail against any defendant (Gottlieb v. Orange County, 1996).

Consider also the decision of the United States Court of Appeals for the Third Circuit (which encompasses Delaware, New Jersey, Pennsylvania, and the Virgin Islands) in Miller v. City of Philadelphia (1999), a case of alleged physical abuse by a mother that led to the temporary removal of her three children. There, in a shoddy investigation, the CPS investigator asked the children leading questions, requested that the mother produce all three children for a physical exam even though the abuse allegation pertained to only one child, met secretly with a hospital social worker, excluded the mother’s attorney from the waiting area outside the examination room, and was advised by a doctor that it was not clear whether the child’s bruises were accidental or a result of physical abuse. Not surprisingly, the caseworker received employment reviews indicating he did not always follow proper procedures. Still, the court concluded that “[e]ven if all of the facts alleged … were true, [the investigator] did not act in a way that shocks the conscience” (p. 377). Clearly, Miller sets a high bar for actionable conduct.

Similarly, in another physical abuse case, county officials acted reasonably when they temporarily removed a 15-year-old daughter based on the fact that her mother pulled her from their car by her hair, wrestled her to the ground, and pushed her face into a gravel driveway, causing minor bruises, cuts and scrapes, and the child arrived at school visibly distressed (Patterson v. Armstrong County Children and Youth Services, 2001).

CPS Misdeeds are Sometimes Overlooked When Others Could Correct Them

Even particularly egregious acts may be insulated from liability where a wronged parent cannot show a connection between the act and the claimed constitutional violation. For instance, one appeals court tossed out a jury verdict for a removed father where he failed to avail himself of opportunities to clarify how long he needed to stay away. In Terry v. Richardson (2003), a 3-year-old girl, Jaidah, returned from visits at her father’s house withdrawn and afraid of other men. When asked by her mother whether she and her father had any “secrets,” Jaidah said yes—at which time her mother, Richelle, contacted Cheryl Richardson, a caseworker (p. 782). Richardson left Jaidah’s father, John Terry, a message the next morning informing him that he should not see or contact Jaidah. When Terry called her back, he seemed to understand Richardson’s reasoning. Two physicians corroborated the existence of sexual abuse, and for the next month and a half Jaidah continued to implicate her father when questioned about the abuse. During this time, Jaidah missed one scheduled visit with Terry because Jaidah was sick.

Richardson interviewed Terry 15 days into the investigation and again advised him not to contact Jaidah until the investigation was complete. On the 48th day, she called Terry to inform him her investigation was complete and that Jaidah’s accusations seemed to be true. Terry denied receiving the message. Richelle then obtained an order prohibiting Terry’s visitation with Jaidah. Subsequently, a court found that Jaidah had been abused, but not by Terry. Terry brought suit against Richardson and a jury awarded him $2,062 and Jaidah $7,210.

The United States Court of Appeals for the Seventh Circuit (which covers Illinois, Indiana, and Wisconsin) reversed the verdict, finding no constitutional rights had been infringed (Terry v. Richardson, 2003). First, Terry had ample opportunity to ask Richardson about the extent of her authority. Although Terry’s attorney spoke with Richardson, he did not ask when Terry could see his daughter. Second, any incursion on Terry’s rights was minor—at most, Richardson prevented Terry from seeing Jaidah for 1 day. Finally, while the court noted that “arbitrary abuses of government power are checked by requiring objective justification for steps taken during the investigation” (p. 787), it found such justification here.

Similarly, in Miller v. City of Philadelphia (1997), a mother who temporarily lost custody of her three children due to allegations of physical abuse, alleged that a child welfare worker attempted to suborn perjury, induced the examining hospital to falsify records, and misrepresented the physician’s medical report to the judge who issued the temporary protection order. Although the trial court initially denied qualified immunity for the caseworker (Miller v. City of Philadelphia, 1997), the Third Circuit, after several rounds of appeals, concluded that “even if [the caseworker] did misrepresent the doctor’s report to [the prosecutor, the mother] failed to establish a causal connection between the alleged misrepresentation and the Judge’s decision to grant a separation order” (Miller v. City of Philadelphia, 1999, p. 374). Although she had ample opportunity to do so, the mother chose not to depose the physician or prosecutor, “both of whom would have had direct knowledge of any misstatements or misdeeds” (p. 374) by the caseworker. Moreover, the prosecutor “spoke independently with [the physician] to ascertain his opinion,” which “should have served to expose any lies” (p. 374). For this reason, “any subsequent misstatements by [the prosecutor] to the Judge during their telephone hearing would not have been caused by” (p. 374) the caseworker. While no one endorses such questionable practices, it is nonetheless instructive that such actions still did not trigger liability.

Like the actions of caseworkers, court orders also enjoy significant deference. “[0]rders of protection are rarely struck down as ‘unreasonable.— Few are appealed, and, when they are, appellate courts tend to rely on the expertise” (Besharov, 2004, n.p.) of the lower court. Protective orders on behalf of sexually abused children have been upheld in numerous cases, even where the order impacts the offending parent’s access to a residence he shared with the child (Campbell v. Campbell, 1991; Cooke v. Naylor, 1990; Kenekerv. Keneker, 1991).

Stepping over the Line

Although courts accord caseworkers significant protection, circumstances exist in which caseworkers can and do exceed the wide latitude given them. A caseworker would be advised not to suborn perjury, induce medical providers to falsify records, or misrepresent a medical report to the presiding judge, as alleged in Miller v. City of Philadelphia (1997, p. 1066). Reckless disregard for the facts is also not prudent. In Croft v. Westmoreland County Children and Youth Services (1997), the court found that a caseworker lacked “objectively reasonable grounds” (p. 1127) when she threatened to remove a child if the father did not exit the home, based only on an anonymous tip passed along a chain of four persons, without any corroboration. The caseworker acknowledged that she renewed the ultimatum even after her interviews with the parties left her with no “opinion one way or the other” (p. 1127) that the father was sexually abusing his son.

Where it is not clear that an objectively reasonable basis existed for acting, courts will also allow litigation to proceed beyond the initial stages—known as summary judgments proceedings. In Puricelli v. Houston (2000), a social worker relied upon uncorroborated anonymous reports of abuse in allegedly issuing an ultimatum to a parent to leave his home. By allowing the case against the social worker to proceed to trial, the court permitted a jury to decide whether the social worker had a reasonable basis for issuing the ultimatum, if that is what occurred (p. 8).

In sum, caseworkers have significant latitude to direct alleged offenders to exit the household, rather than immediately proceeding to the usual remedy of removing the victim and other children from their home. This latitude, together with the explicit legal authority granted to them in a number of states, should embolden them to do the obvious—remove the threat from the household rather than the children.

The Case for Removal

A strong case can be made for excluding alleged offenders and leaving the children in place. There are compelling reasons for taking this approach.

The Consequences for Children of Removal

By excluding the alleged perpetrator, the home becomes a safer environment not only for the victim, but also for every child in the house (Cross et al, 1999, p. 41; Wilson, 2002). Exclusion offers benefits in addition to safety. The support a child receives from her non-offending mother moderates the long-term effects of the abuse itself (Everson, Hunter, Runyan, Edelsohn, & Coulter, 1989). Exclusion, unlike removal, offers the child the possibility of such support. A child who has endured abuse at the hands of an adult should not then be subjected to the “double victimization” (Bagley & King, 1990, p. 101) of “system-induced trauma” (MacFarlane & Bulkley, 1982, p. 72) that forces him or her to leave familiar surroundings and the comfort of his or her mother and siblings. This trauma can be considerable.

A removed child is often cut off from all contact with his or her non-abusing mother for extended periods of time (Levy, 1989). He or she may “develop feelings of guilt or unworthiness, especially if [he or she] was the one to disclose the abuse” (Ryan et al., 1991, p. 125). While not every removed child is fostered, those who are placed in foster care may experience serious psychological damage (Wald, 1975, pp. 993-994).

Sometimes removal “places a child in a more detrimental situation than he would be in without intervention” (Wald, 1975, pp. 993-994). A child may be sexually victimized in foster care. A 1999 study found that foster care was a significant risk factor for sexual abuse and that foster parents were the perpetrator nearly one-third of the time (Hobbs, Hobbs, & Wynne, 1999). In another study, foster fathers and other foster family members were the perpetrators in over two-thirds of the substantiated cases (Benedict & Zuravin, 1996). Physical abuse may also occur (Gelles & Cornell, 1990).

In many instances, the child’s abuse at the hands of a foster parent comes as no surprise to the state. Rosenthal, Motz, Edmonson, and Groze (1991) found that reports of child sexual abuse while in an out-of-home placement—defined to include family foster care, group homes, residential treatment, and institutions—were the most likely to be confirmed and that in 27% of all maltreatment reports, prior allegations against the perpetrator were present. As Gelles (1996) noted, “in some cases, foster parents are actually more dangerous to the child than the biological parents are” (p. 162).

Even where a child is not directly victimized, removal can be a bad idea. Separating the child from his or her mother frustrates the “laborious task of putting lives back together” (Herman & Hirschman, 1981, p. 144). The “essential nucleus” for this healing process is the mother-child relationship (p. 144). Removal also exposes the child to a litany of ills caused by “foster care drift” (Goldstein, 1999, p. 714). The extent of this dislocation cannot be understated. In one study, 13% of sexually abused children placed in foster care experienced six or more moves during their time in foster care (Bolen, 2001, p. 229).

More fundamentally, disrupting the parent’s life, rather than the child’s, seems preferable where the allegations initially appear to be true. As one court noted in a domestic violence case, “[a] victim of… outrageous and life-threatening sort of abuse … cannot be held hostage to the potential homelessness of her abuser, who created the intolerable situation in the first instance” (V.C. v. B.C., 1999, p. 453). The equities are especially compelling where the allegations turn out to be substantiated. In that instance, “the father … is responsible for the choice to eroticize [his] relationship with [his child]” (Salter, 1988, p. 42) and so should bear the consequences of that choice even when he is not prosecuted. To do otherwise permits offenders to externalize the cost of their behavior on their victims who, ironically, are removed to ensure their safety.

Guam and Hawaii both essentially take this approach. In Guam, the court must first give “due consideration to ordering the removal… of the alleged perpetrator from the child’s family home” (19 Guam Code Ann. § 13316, 2004) before removing the child. In Guam and Hawaii, the child’s family bears the “burden of establishing that it is not in the best interests of the child that the alleged perpetrator be removed from the family’s home” (19 Guam Code Ann. S 13316; Haw. Rev. Stat. Ann. § 587-53(f), 1999). Texas errs on the side of the child even more forcefully. There, if the state CPS agency files a petition to exclude the alleged offender, the court has no choice but to exclude the parent if it finds that the child has been sexually abused and “there is substantial risk” he or she will be abused again if the parent is not excluded from the home (Morgan & Gaither, 1999).

While novel in some jurisdictions, excluding an accused parent is certainly no more radical than what we do in private disputes between adults at the time of divorce. Courts routinely direct one spouse to leave the home (which is tantamount to awarding exclusive possession of the marital home to one party) (Jetter v. Jetter, 1971). And, of course, we remove children every day without thinking twice about the considerable power being wielded by the state (American Prosecutors Research Institute, 2004).

Importantly, the government acts preemptively before criminal adjudications in other contexts. Bond hearings commonly “place restrictions on … place of abode of the person during the period of release” when that person poses an “unreasonable danger to the community” (18 U.S.C. § 3142(c), 2000; S.C. Code Ann. § 17-15-10, 2003). All jurisdictions in the United States take this into consideration (8A Am. Jur. 2d Bail and Recognizance § 34, 1997). Literally thousands of times each day, judges place restrictions on persons presumed innocent. It is true that a bond follows arrest but, as with an allegation of abuse, there has been no hearing on the merits or conviction.

Addressing Valid Process Concerns

There is no doubt that removal of the alleged offender will sometimes raise significant due process concerns. There is some evidence, in fact, that caseworkers have used “voluntary” agreements to leave the home as a means of short-circuiting the normal protections built into the CPS system. Pearson (1998) notes that “authorities sometimes employ coercive tactics… as an avoidance of procedural safeguards for the handling of child abuse investigations” (pp. 842-843). This behavior cannot be condoned.

The solution to such overreaching, however, is not to take this remedy out of the state’s arsenal. Instead, we should institutionalize and heavily regulate it, as more than a half-dozen states now do. Maine extends the same process protections to parents who are asked to exit as it does when pursuing other equally drastic remedies, like the removal of children from their homes (Me. Rev. Stat. Ann. tit. 22, § 4036(1)(F-1), 1998). These protections include providing legal counsel for the parent, a guardian ad litem for the child (Me. Rev. Stat. Ann. tit. 22, § 4005, 1998), notice and opportunity to participate in a hearing (§ 4006), and, where the order was issued on an emergency basis, a preliminary hearing within 21 days (§ 4006).

Texas requires notice, sets a 14-day outer limit for any temporary restraining order, and erects a four-part test that must be satisfied before a temporary restraining order can be issued. Among other things, the state must show that there “is no time, consistent with the physical health or safety of the child, for an adversary hearing” (Morgan & Gaither, 1999). Kentucky courts instruct judges who order alleged perpetrators to “stay out of the family home” (Ky. R. Jefferson Fam. Ct. Rule 6 app., 2003) to do so with great specificity, defining the specific distance the person should stay away. Protective orders in New York must be for a specified time period, not to exceed a year initially, unless certain aggravating circumstances exist (N.Y. Fam. Ct. Act § 842, 2004). These protections balance the state’s interests in quick but accurate adjudications with the alleged offender’s interests in the least restrictive remedy, and the child’s interests in not being cleaved from the security of their family and home. Most important, they reduce significantly legitimate concerns about possible overreaching by CPS caseworkers.

Limitations to Excluding Alleged Offenders

Excluding alleged offenders is not without its limitations. Just as a child who is removed from his or her home may experience guilt, so may a child whose parent is ejected, especially when the “family suffers economically” (Ryan et al, 1991, p. 125). In addition, like the decision to remove a child, the decision to exclude an alleged offender is made “against a background of urgency and inadequate information” (Pickett & Maton, 1977, p. 63) and will sometimes turn out to have been unwarranted. The fact that an allegation may later prove unfounded should not, by itself, dissuade us from using this remedy, however. These error costs are no different than those that occur when the state removes a child who is later found not to have been abused.

The real “difficulty with restraining orders is that they are hard to enforce and, in the case of child sexual assault, depend upon the presence of an adult ally for the child to monitor the situation and to report any violation of the restraining order” (Graves & Sgroi, 1982, p. 328). Clearly, it is essential that the non-abusing parent be supportive. In Great Britain, an accused parent may not be excluded from the household during the investigation if another person in the home is not willing to care for the child. The other adult must also consent to the exclusion (Children Act 1989, ñ 41 § 38A (3)). For restraining orders to be available in Texas, the court must find that the child “is not in danger of abuse from a parent… with whom the child will continue to reside” (Morgan & Gaither, 1999). The remaining parent must “make a reasonable effort to monitor the residence” and agree to report any attempts by the excluded parent to return home. The failure to do these things is a misdemeanor, as is the perpetrator’s return to the residence (if the perpetrator has been previously convicted of returning, the return constitutes a felony) (Morgan & Gaither, 1999). Strong empirical evidence demonstrates that most non-offending mothers are capable of performing this critical safety function, and they should be trusted to do so unless screening tools suggest that an individual mother cannot or will not do so.

Conclusion

Nearly everyone recognizes that “[w]e need to develop alternatives to prosecution that can increase children’s safety without making them leave their homes” (Cross et al., 1999, p. 43). The easiest, most direct route to this is to take the alleged offender out of the home, not the children. Although the perceived “inability to remove the offender” (Bolen, 2002, p. 58) remains strong, we have come a long way since Florence Rush asked in 1974, “Has anyone thought of the fantastic notion of getting rid of the father?” (p. 71). Over the last quarter century, judges and members of the legislature have done the heavy lifting so that removing alleged offenders is no longer unthinkable. And when practice finally catches up to the law, fewer children will needlessly endure the horror of incest and the trauma of being ripped from the security of their homes.