sign in sign up
<<
Home , Be (Cyrillic), Child abuse, Crime drop, Sexual abuse, Violence

U.S. Department of Office of Justice Programs Office of Juvenile Justice and Delinquency Prevention

J. Robert Flores, Administrator January 2004

Explanations for Message From OJJDP Decline in The decline in sexual cases, as reported by child protective serv- Cases ice agencies, could be an encourag- ing development if future analysis and research establish its signifi- cance. The uncertainty about the and Lisa M. Jones meaning of this trend, however, underscores how much remains to The Office of Juvenile Justice and Delinquency Prevention (OJJDP) is committed to be accomplished in terms of drawing improving the justice system’s response to against children. OJJDP recognizes lessons from past experience to deter- that children are at increased risk for victimization. Not only are children the vic- mine future policy. Clearly, a trend of tims of many of the same crimes that victimize adults, they are subject to other crimes, this magnitude merits further inquiry like and , that are specific to childhood. The impact of these crimes and analysis. on young victims can be devastating, and the violent or sexual victimization of children For example, it would be useful to can often to an intergenerational cycle of and abuse. The purpose of know whether policies of more ag- OJJDP’s Crimes Against Children Series is to improve and expand the nation’s efforts gressive prosecution, incarceration, and treatment have played a role. to better serve child victims by presenting the latest information about child victimization, If so, what will be the impact when including analyses of crime victimization statistics, studies of child victims and their spe- many of the large group of sexual cial needs, and descriptions of programs and approaches that address these needs. abuse offenders placed in custody in recent years are released from their sentences? From another per- The number of sexual abuse cases sub- Identifying the source or sources of the spective, have efforts to educate stantiated by child protective service decline in the number of substantiated children and to identify and treat ju- (CPS) agencies dropped a remarkable 40 sexual abuse cases is important. The venile sex offenders had the effect percent between 1992 and 2000, from an possibility that a real decline occurred is of reducing the number of victims estimated 150,000 cases to 89,500 cases, heartening and could point the way to and perpetrators? but professional opinion is divided about more effective strategies for preventing This Bulletin reviews six plausible why (Jones and Finkelhor, 2001; Jones, all kinds of child maltreatment. On the explanations for the decline in sex- Finkelhor, and Kopiec, 2001). It is possible other hand, if the decline is due solely ual abuse cases in light of available that the incidence of sexual abuse has to decreased reporting or changes in CPS data. While acknowledging the like- declined as a result of two decades of procedures, it could mean that more chil- lihood that multiple factors are in- prevention, treatment, and aggressive dren are failing to get the help and serv- volved, the authors find activity. It is also possible ices they need. of a significant decline in child sex- that there has been no real decline, and ual abuse. that the apparent decline is explained by This Bulletin explores the strengths and adrop in the number of cases being iden- weaknesses of six possible explanations Federal, state, and local agencies tified and reported or by changes in prac- for the decline by using data from a num- should collaborate on a research tices of agencies. ber of different sources (see page 3): agenda that will cast light on the aggregate data from the National Child meaning of this trend.

Access OJJDP publications online at www.ojp.usdoj.gov/ojjdp Abuse and Neglect Data System (NCANDS); at more comparable numbers, sexual decline in (15 percent) oc- detailed child protective service data from abuse totals were extrapolated to account curred between 1998 and 1999, whereas the Illinois, Minnesota, Oregon, and Pennsylva- for the population of all 50 states and the more gradual 40-percent decline for sexual nia; and self-report data from the National District of Columbia as estimated annually abuse occurred over an 8-year period. Crime Victimization Survey (NCVS) and by the .S. Census. These extrapolated from schoolchildren in Minnesota. It pro- totals show that the number of substan- vides substantially more evidence about tiated sexual abuse cases reached a peak Explanations for the the decline than was available in a previ- of approximately 149,800 in 1992, followed Decline ous Bulletin on the same topic, The - by declines of 2 to 11 percent each year As part of their earlier research on this cline in Child Sexual Abuse Cases (Jones through 2000, the last year for which data topic, the authors conducted a survey of and Finkelhor, 2001). are available (see figure 1). In 2000, esti- state child protection administrators to mated cases of sexual abuse reached a low gather hypotheses and evidence about the of approximately 89,355, for a total decline Key Findings decline in sexual abuse (Jones, Finkelhor, of 40 percent in identified sexual abuse and Kopiec, 2001). Although the adminis- ◆ Detailed data provided by four state cases over the 8-year period. The trend is trators expressed many ideas about the CPS agencies offered little evidence not universal, but it has occurred in the decline, six explanations were offered fre- that the decline was due either to more majority of states. Of 49 states, 39 experi- quently and backed by some anecdotal conservative judgment by CPS about enced a total decline of 30 percent or more support: the types of sexual abuse cases they in substantiated cases of sexual abuse would investigate or substantiate or from their peak year to 2000, and 19 of ◆ Increasing conservatism within CPS. to increasing reluctance by CPS to these states saw declines of more than 50 In this view, sexual abuse cases were become involved in cases in which the percent in their sexual abuse caseloads. declining in state caseloads because perpetrator is not a primary caregiver. CPS was adopting more conservative The decline in sexual abuse does not ap- standards regarding “questionable” ◆ There also was no strong evidence that pear to be just an extension of a general cases (.g., allegations arising in di- the decline was largely due to a dimin- declining trend in overall child maltreat- vorces and custody disputes) or ishing reservoir of older, ongoing cases ment or of some other demographic fac- cases with weak initial evidence. available for new disclosures. tor. According to estimates based on the ◆ Exclusion of cases that do not involve ◆ There was some evidence that the sex- NCANDS data, the decline in sexual abuse caretakers. In this view, CPS was in- ual abuse decline in one state could be appears to account for a large part of the creasingly excluding from its jurisdic- partly explained by changes in CPS pro- 15-percent decline in child maltreatment. tion sexual abuse cases in which the cedures and data collection methods. Neglect cases have fluctuated during the perpetrator was not a primary caregiver. According to national data, however, 1990s with no overall decline while physi- this explanation does not successfully cal abuse has declined 30 percent since ◆ Changes in CPS data collection meth- account for the declines seen in the a peak in 1995. The decline in physical ods or definitions. In this view, the majority of states. abuse is significant, but it is smaller and decline was due to changes in the way more recent than the decline in sexual CPS tabulated or counted its cases, ◆ There was mixed evidence that report- abuse. In fact, the largest proportion of the such as changing from a three-tiered ing of sexual abuse to CPS declined because of a “backlash,” that is, a greater public and professional skepti- cism about reports of sexual abuse. Figure 1: Estimated Number of Substantiated Cases of Sexual Abuse ◆ Evidence from a number of different in the , 1990Ð2000 sources, including NCVS data showing a 56-percent decline in self-reported sexu- 160,000 al against juveniles, is consis- tent with a real decline in sexual abuse. 150,000 ◆ Finally, additional studies and improved 140,000 40% decline data are needed to make crucially im- 130,000 portant decisions for public policy based on the factors that are most 120,000 responsible for the decline. 110,000

Number of Cases 100,000 Evidence for the 90,000 Decline 80,000 Yearly estimates of substantiated sexual 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 abuse from 1992 to 2000 were calculated from CPS administrative data collected by a r NCANDS. The number of states that sub- mitted data to this system each year var- Source: Authors’ analyses of data from 1990Ð2000 National Child Abuse and Neglect Data ied from 43 to 49, making published totals System (NCANDS) reports (U.S. Department of and Services, 1992Ð2002). difficult to compare across years. To arrive

2 Data Sources on Sexual Abuse Trends National Child Abuse and Neglect Data System (NCANDS). NCANDS, organized by the U.S. Department of Health and Human Services’ Children’s Bureau, collects information from state child welfare agencies on a number of variables related to child maltreatment. Data collection through this system began in 1990. Although a growing number of states submit case-level data to NCANDS, aggregate data from the Summary Data Component (SDC) were used to calculate national sexual abuse trends from 1990 through 2000. The SDC provides annual estimates of child maltreatment cases substantiated by child protec- tive services. These data have many limitations, including the fact that states vary considerably in how they define maltreatment and how they investigate and count cases. State Child Protective Service Data. To examine sexual abuse trends in more detail, the authors also examined state CPS data from Illinois, Minnesota, Oregon, and Pennsylvania. These states showed large declines in substantiated sexual abuse and were among the few states for which consistent and extensive case-level CPS data were available from the early 1990s through 1999. The data provided by these states allowed an examination of trends in sexual abuse investigations by the age of the victim, the type of abuse, the perpetrator’s relationship to the victim, the age of the perpetrator, and the investigation outcome, among other factors. National Crime Victimization Survey (NCVS). NCVS is conducted annually by the U.S. Department of Commerce’s Bureau of the Census on behalf of the U.S. Department of Justice’s Bureau of Justice Statistics. Approximately 55,000 U.S. households with a total of 100,000 individuals ages 12 and older are surveyed each year. The survey collects information about the charac- teristics of victimizations, including victim and perpetrator demographics, the incident location, and a description of the incident. Minnesota Student Survey. The Minnesota Student Survey is a voluntary, anonymous, self-administered questionnaire that asks students about a range of experiences including substance use, sexual behavior, and . Two survey ques- tions ask about sexual abuse victimization. The survey has been administered to 6th, 9th, and 12th grade students in Minneso- ta five times: in 1989, 1992, 1995, 1998, and 2001. Approximately 90–99 percent of Minnesota’s school districts have participat- ed in the survey each year, involving more than 100,000 students. For trend analyses, data are limited to the approximately 69 percent of Minnesota’s school districts that participated in the survey in all 5 years. A weighting procedure was used to adjust for differences in student participation rates across districts.1

1 For more information about the Minnesota Student Survey’s methodology, see Harrison, Fulkerson, and Beebe (1997), or Minnesota Department of Children, & Learning (2001).

classification system (substantiated/ Explanations and the is, if CPS workers were applying more indicated/unsubstantiated) to a conservative standards and confirming two-tiered system (substantiated/ Evidence less sexual abuse, there might be a greater unsubstantiated). The following sections review evidence drop in substantiated than in reported that might be marshaled to support each ◆ Less reporting to CPS due to a sexual or investigated cases. of the six possible explanations for the abuse backlash. In this view, negative reduction in sexual abuse cases. Evidence However, there is not much evidence for publicity about sexual abuse cases and was drawn from a variety of sources but such a pattern. For example, in Illinois, the potential liability of professionals especially from the four states with exten- substantiations and investigations de- who report suspected abuse made the sive CPS data. clined in tandem (see figure 2). Similar public and professionals more reluc- patterns were identified in Minnesota and tant to report sexual abuse. Pennsylvania. Nationwide, evidence pro- ◆ Increasing Conservatism vided by the 50-state survey shows that A diminishing reservoir of older Within CPS cases. In this view, there had been a the substantiation rate for all types of reduction in the supply of older but One very plausible explanation of the maltreatment remained constant at 33 previously undisclosed cases available decline in substantiated sexual abuse percent from 1994 through 1999 (Peddle for new disclosures but no true decline cases is that child protection agencies and Wang, 2001). in new cases. have become more conservative regarding the cases they investigate or substantiate. If increasing conservatism among CPS ◆ A real decline in the incidence of This may reflect a better knowledge base, staff explained the decline, then a greater sexual abuse. In this view, there was a reaction to about overzealous number of conservative judgments would areduction in the number of children investigations, caution in anticipation of also be expected in certain kinds of poten- actually being abused as a result of legal retaliation, an effort to triage cases tially questionable cases. In recent years, increased prevention efforts, more brought about by limited resources, or a skeptics concerned about unnecessary prosecution and incarceration of concern about unnecessarily stigmatizing investigations have particularly targeted offenders, or other social or cultural people by subjecting them to investiga- certain allegations, such as those arising 1 changes. tions based on questionable information. in and custody disputes; those An obvious place to look for increasing involving very young children, whose skepticism about questionable cases is often deemed unreliable; 1 Specific mechanisms for a real decline will be and those involving less intrusive forms explored in subsequent research. would be in the substantiation rate. That

3 that the sexual abuse allegation was an Figure 2: Illinois Trends in Sexual Abuse Investigation and to manipulate an already existing Substantiation, 1990Ð99 custody conflict.) Sexual abuse allegations arising in previously existing custody dis- putes would be expected more often when 25,000 children reside in single- or step- 20,000 parent households than when they reside 38% decline in households with two biological . 15,000 Agreater decline could be expected in such cases if CPS became increasingly 10,000 skeptical about them. Trends in Pennsyl- 43% decline vania do not follow that pattern. In fact, 5,000 figure 3 shows that the largest declines in Number of Cases Pennsylvania have occurred in allegations 0 coming from families with two biological 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 or adoptive parents, instead of families Ye a r with single parents or stepparents. Trend data are somewhat more support- Investigated Substantiated ive of the hypothesis that increasing con- Source: Authors’ analyses of unpublished data provided by the Illinois Department of servatism among CPS staff is affecting the Child and Services. decline in substantiated cases of sexual abuse for cases involving younger chil- dren. In Illinois, Minnesota, and Pennsyl- vania, the decline was most pronounced Figure 3: Pennsylvania Trends in Sexual Abuse Cases, by Living for cases involving children younger than 6. For example, figure 4 shows a 57- Situation of Victim at Time of Report percent decline in Illinois cases with child victims ages 0–2, compared with an aver- 9,000 age 26-percent decline across other age Percent Decline groups. However, the decline has not been 7,500 39% limited to cases with victims ages 0–2; 6,000 it has occurred across all age groups, with 31% a higher rate of decline for cases involving 4,500 very young victims. Because the very 22% young make up a small portion of all vic- 3,000 tims of sexual abuse, even fairly big de- clines for this age group are not enough Number of Cases 22% 1,500 to account fully for the overall decline in 12% sexual abuse cases. 0 1990Ð92 Average 1997Ð99 Average Where there is other information from states on the quality or seriousness of the One biological or and Other Two biological allegation or evidence, it generally does or adoptive adoptive parent parent parent subtitute parents and one stepparent not support the hypothesis that there were especially dramatic declines in Source: Authors’ analyses of unpublished data provided by the Pennsylvania Office of questionable cases. Case records in Children, , and Families. Pennsylvania record the presence of med- ical evidence or perpetrators’ confessions. As shown in figure 5, the decline in Penn- of sexual . If increasing conser- 3–6. Unfortunately, case reports in the sylvania cases with such strong evidence vatism about questionable cases were an available states do not indicate whether has been as great, if not greater, than the important factor in the decline of sexual they were preceded by a custody dispute. decline in weaker cases. Illinois records abuse, it would be reasonable to predict In Pennsylvania, however, case reports do the presence of sexually transmitted a differentially greater decline in these indicate the family structure of the child’s diseases (STDs) and allegations of pene- types of cases. At the same time, a less household. Cases in which a custody tration. Here again, the declines are sub- pronounced decline might be seen in dispute precedes a sexual abuse allega- stantial in cases in both categories, and cases involving more credible evidence, tion are less common in families where they are lowest in the category of ex- such as medical evidence or a perpetrator two biological parents are present at the ploitation alone, which might contain confession, and more serious kinds of time of the allegation. (Sometimes a cus- some of the least serious cases and those contact, such as penetration. tody dispute follows the allegation of sex- with the weakest evidence (see figure 6). ual abuse in an intact family, but in these Trend data for cases involving question- The finding that the decline in sexual cases, CPS would be less likely to assume able allegations are presented in figures abuse cases with stronger evidence was

4 comparable to that of cases without such evidence is far from fatal to the “increas- Figure 4: Illinois Trends in Sexual Abuse Cases, by Age of Victim ing conservatism” argument. For example, at Time of Report large declines in STDs may be due to the decline in STDs in the general population 20,000 (Centers for Disease Control and Preven- Percent Decline tion, 2000). Standards of medical evidence 16,000 21% may also have become more conservative 31% (Carole Jenny, Brown University, personal 12,000 communication, May 14, 1997), and per- 33% petrators may be newly emboldened by 8,000 attorneys not to confess. But the 29% data do not give strong or consistent evi- 4,000 37% dence that the exclusion of questionable Number of Cases cases—those that involve controversial 57% 0 allegations or weaker evidence—is a major 1990Ð92 Average 1997Ð99 Average factor in the decline in sexual abuse cases. 15 or older 12Ð14 9Ð11 6Ð8 3Ð5 0Ð2 Exclusion of Cases That Do Not Involve Caretakers Source: Authors’ analyses of unpublished data provided by the Illinois Department of Child and Family Services. According to child protection officials, in recent years, some states have dealt with burgeoning caseloads and declining resources by narrowing the scope of child welfare responsibilities (Jones, Finkelhor, Figure 5: Pennsylvania Trends in Indicated Cases of Sexual Abuse, and Kopiec, 2001). Because the traditional by Type of Evidence and statutory responsibilities of child pro- tection have concerned threats to chil- 4,000 dren’s well-being at the hands of caretak- Percent Decline ers, some state agencies have increasingly 37% excluded cases that involve allegations 3,000 of abuse by noncaretakers, including non- relatives, distant relatives, and juvenile 58% offenders. Compared with other kinds of 2,000 maltreatment, sexual abuse cases have typically involved more allegations of of- fenses by noncaretakers. If a rise of such

Number of Cases 1,000 37% exclusionary policies were responsible for the decline in substantiated cases of sexual abuse, greater declines could be 0 expected in cases involving perpetrators 1990Ð92 Average 1997Ð99 Average who are not caretakers. Medical Perpetrator Evidence from The available data from the states exam- evidence admission CPS investigation ined for this Bulletin show some evidence of a differential decline for cases that in- Source: Authors’ analyses of unpublished data provided by the Pennsylvania Office of volve noncaretakers and juvenile perpe- Children, Youth, and Families. trators, but this evidence was not strong. There may have been some differential decline in cases involving young perpetra- exact opposite of what the explanation expected in cases involving nonfamily or tors. Trend data for Pennsylvania showed would have predicted (see figure 8). more distant perpetrators because a a larger decline in cases with perpetra- classification of neglect would apply tors ages 12–17 than in cases with older The same data are relevant to another, primarily in such cases. CPS workers are perpetrators (see figure 7). However, the related explanation for the decline in sex- not likely to label a case of sexual abuse number of teen perpetrators was already ual abuse cases: namely, that the decline by a father or stepfather simply as neglect small, and the larger percentage of decline is due to an increase in CPS labeling of on the part of the . The more dra- did not account for much of the overall sexual abuse cases as neglect or “failure matic decline in parental perpetrators decline. There was no evidence for the to protect” on the part of the nonperpe- seems to weigh against both the “failure exclusion of cases with noncaretaker trating caretaker. That practice may have to protect” and the noncaretaker exclu- perpetrators in the three states where developed as CPS agencies more carefully sion explanation in the states that were these data were available. For example, in defined their role as ensuring the protec- examined. Illinois, the decline was most dramatic for tion of the child in the home. However, cases involving parental perpetrators, the here again, a larger decline would be

5 systems as possible causes for the de- Figure 6: Illinois Trends in Sexual Abuse Cases, by Type of cline. These changes might, in principle, Evidence/Allegation be disruptive to data collection, resulting in artificial declines.

20,000 Percent Decline If such changes were behind the decline, state trends could be expected to show a 46% significant decline occurring over a short 16,000 period of time paralleling the implemen- tation of these changes. An examination 32% of state trends using NCANDS data (see 12,000 the sidebar on page 3) identified 15 states in which a very large decline (more than 33 percent) in sexual abuse cases occurred 8,000 37% in 1 year. When state officials from these “discontinuous decline” states were asked Number of Cases 4,000 about these trends, more than half were able to explain a change in procedure 21% that could account for the trend. An exam- 0 ple of this effect was evident in the CPS 1990Ð92 Average 1997Ð99 Average data from Oregon, which experienced a 60-percent decline in substantiated cases Sexually transmitted Penetration Molestation Exploitation (see figure 9), but much of the decline oc- diseases curred between 1993 and 1994. Discus- sions with Oregon state officials indicated Source: Authors’ analyses of unpublished data provided by the Illinois Department of that prior to 1994, a determination about Child and Family Services. whether abuse occurred was made with every report, making reports indistin- guishable in the database from face-to-face investigations. This policy was changed Figure 7: Pennsylvania Trends in Sexual Abuse Cases, by Age in 1994, and the result was a drop in the of Perpetrator number of cases that were labeled as investigated and substantiated. 9,000 Such changes in data collection proce- Percent Decline dures or definitions are unlikely, however, 7,500 to account for the large and gradual de- 6,000 22% clines in child sexual abuse cases that occurred in the majority of states. Fur- 4,500 thermore, if changes in definition and data 32% collection procedures were occurring, 3,000 state trends would be expected to show

Number of Cases 33% both increases and decreases in the data. 1,500 40% Changes in definition or data collection do 0 not acceptably explain the size and con- 1990Ð92 Average 1997Ð99 Average sistency of the national decline.

35 and older 25Ð34 18Ð24 12Ð17 Less Reporting to CPS Due Source: Authors’ analyses of unpublished data provided by the Pennsylvania Office of Children, to a Sexual Abuse Backlash Youth, and Families. Another frequently voiced explanation of the decline in child sexual abuse is that professionals who would report abuse Changes in CPS Data these changes were artificially creating have been intimidated by a “sexual abuse Collection Methods or what appeared to be a declining trend backlash” resulting from negative publicity in the media about false allegations, over- Definitions (Jones, Finkelhor, and Kopiec, 2001). For example, some officials reported that their zealous reporting, and lawsuits seeking During the 1990s, a number of state child state had moved from a three-tiered classi- damages against those who make reports. protection agencies changed the way they fication system (substantiated/indicated/ Starting in the mid-1980s, highly publicized define and categorize child maltreatment unsubstantiated) to a two-tiered system cases in Jordan, MN, Wenatchee, WA, and and how they collect and enter informa- (substantiated/unsubstantiated). Others Manhattan Beach, CA, raised questions tion about the cases that they investigate. described changes such as the establish- about whether authorities were unfairly In a survey conducted by the authors, ment of central registries and the imple- targeting innocent citizens. New organiza- many child protection officials felt that mentation of more detailed data collection tions, advocates, and legal experts came

6 practice are an example of the latter Figure 8: Illinois Trends in Sexual Abuse Cases, by Perpetrator’s group. Reporting child abuse has always Relationship to Victim posed a burden and a risk for physicians because it potentially alienates patients. As independent practitioners, they may 20,000 Percent Decline also feel particularly vulnerable to dam- age claims. 16,000 39% The state data that were examined for this 12,000 Bulletin did not, in any instance, include evidence of a rise in anonymous reporting. 8,000 29% In Illinois, however, anonymous reports 18% declined less than the average rate, which

Number of Cases 4,000 31% could reflect some shift in preference 28% toward anonymous reporting, but within 0 the overall context of a decline in reports 1990Ð92 Average 1997Ð99 Average from virtually all sources (see figure 10). This pattern is not shared by Pennsylva- Parent Other relative Professional Other nia, where anonymous reports declined figure caregiver more, not less, than the average rate.

Source: Authors’ analyses of unpublished data provided by the Illinois Department of The pattern of reporting by physicians in Child and Family Services. private practice is also mixed. In Illinois, reports emanating from such physicians declined much more dramatically than average (see figure 11). Their reports did not, however, decline much more than Figure 9: Oregon Trends in Sexual Abuse Reporting and reports from hospital-based physicians, Substantiation, 1989Ð99 who presumably are more insulated from both economic and legal retaliation by 10,000 families about whom reports are made. In contrast to Illinois, the decline in Pennsyl- 8,000 Ð5% vania reports from physicians in private practice was dramatically less than the 6,000 average and less than the decline in Ð55% reports from hospital-based physicians, 4,000 the opposite of what the backlash expla- Ð60% nation might predict. 2,000

Number of Cases Thus, in Illinois, the evidence is somewhat 0 consistent with the backlash or intimida- 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 tion explanation, although there was a slow- decline and not a true rise in anonymous Ye a r reports. In Pennsylvania, neither anony- mous reports nor reports from physi- Reports Investigated cases Substantiated cases* cians in private practice show evidence of a backlash. Unfortunately, however, be- Source: Authors’ analyses of unpublished data provided by the Oregon Department of cause the backlash explanation concerns Human Services. attitudes and practices among reporters, * Data for trends in substantiated cases were drawn from the National Child Abuse and Neglect it is harder to evaluate with CPS data Data System (NCANDS) annual reports (U.S. Department of Health and Human Services, alone, and the data reported here permit 1992Ð2002). only a weak evaluation of the backlash explanation. together to criticize the criminal justice A backlash or some other fac- A Diminishing Reservoir and child welfare systems’ handling of sex- tor might be expected to affect reports of of Older Cases ual abuse cases (Hechler, 1988). Sources child sexual abuse in two ways: more re- have documented a more negative and ports might be made anonymously, allow- Another possible explanation for the de- critical tone in media coverage of sexual ing reporters to seek protection for a child cline could involve what might be de- abuse and CPS, which may have dissuad- while insulating themselves from criticism scribed as the depletion of the reservoir ed both laypersons and professionals or legal action, and the decline in reports of older, but previously undisclosed, cases mandated to report abuse from reporting might be greatest among those most sus- with no true decline in the number of new (Myers, 1994). ceptible to retaliation, legal action, or cases. The increasing awareness of sexual economic boycott. Physicians in private abuse in the 1980s may have created an

7 artificial surge in the number of cases by There are several predictions that would The survey includes two questions about flushing into view a large number of older be consistent with a true reduction in the experiences with sexual abuse. For sexual or ongoing cases that had previously number of children and youth being sexu- abuse by both family and nonfamily perpe- escaped detection. Only after the reser- ally abused: trators, these data show a slight rise be- voir of old cases was depleted would new tween 1989 and 1992 and a 22-percent ◆ Decline in the number of self-reports disclosures of sexual abuse truly reflect drop from 1992 to 2001 (see figure 13). of sexual abuse by victims. new cases. That is, the influx and subse- This trend also parallels the trend in the quent depletion of old cases would cause ◆ Decline in related social problems. CPS data. sexual abuse rates to decline without any ◆ Greater decline in the most readily true decline in actual incidence levels. Decline in related social problems. If sexu- preventable cases. al abuse were truly declining, the decrease If reservoir depletion is responsible for ◆ Increase in the incarceration of might be paralleled by drops in indicators a considerable amount of the decline in offenders. of other related social problems. These child sexual abuse, then the number of problems could be considered precursors Decline in the number of self-reports of cases involving older children should also or outcomes of sexual abuse, or they could sexual abuse by victims. Unfortunately, show a decline. The victims in older, on- be affected by similar causal factors. The sexual abuse is not a crime category going, and previously undisclosed cases period in which the decline in sexual abuse tracked by the nation’s most reliable would almost certainly be older at the occurred also saw declines in a number of measure of self-reported crime victimiza- time of disclosure because they would other child welfare problems, including: tion, the National Crime Victimization Sur- have been abused for a longer period. vey. However, NCVS does ask about ◆ Figure 4, which shows trends by age of Crime and . and for victims ages 12 victim, demonstrates no differential de- ◆ Births to teenage . and older, and these crimes include acts cline for older victims (see page 5). Thus, counted within the broader definition of ◆ Children running away. there is no evidence for the reservoir child sexual abuse. The NCVS data show ◆ depletion hypothesis. Children living in . that sex offenses against juveniles (ages ◆ 12–17) declined 56 percent between 1993 Teen . A Real Decline in the and 2000, with virtually all the decline In general, the evidence for these other Incidence of Sexual Abuse occurring in offenses committed by declines is more reliable than the evi- If the decline in child sexual abuse cannot known (family and acquaintance) perpe- dence for the decline in sexual abuse. be fully explained by the hypotheses con- trators (down 72 percent, see figure 12). sidered above, then is the decline real? Cases involving known perpetrators are The decline in crime and violent crime the type most likely to be categorized as during the 1990s has been widely publi- sexual abuse. The timing and magnitude cized. The evidence for that decline is based both on self-reports from NCVS Figure 10: Trends in of this decline in self-reports are parallel to the trend in CPS data on sexual abuse. and on reports. NCVS shows a Anonymous Reports of 46-percent decrease in violent crime from Sexual Abuse in Illinois and Another source of self-report information 1994 to 2000 (Rennison, 2001), and a Pennsylvania, 1991Ð98 on sexual abuse comes from the Minnesota 21-percent decrease in intimate partner Student Survey (see the sidebar on page 3). assault from 1993 to 1998 (Rennison and 0

Ð5 Figure 11: Trends in Sexual Abuse Cases Reported by Private and Ð10 Hospital Physicians for Illinois and Pennsylvania, 1991Ð98 Ð15 0 Ð20 -10 Ð25 -20 ercent Change in Number of Cases P Ð30 -30 -40 Ð35 -50 Illinois Pennsylvania -60 ercent Change in Number of Cases

Anonymous Average of P -70 reports reports from all sources -80 Illinois Pennsylvania Source: Authors’ analyses of unpublished data provided by the Illinois Department Private physicians Hospital physicians Average of reports of Child and Family Services and the from all sources Pennsylvania Office of Children, Youth, and Families. The comparison is between Source: Authors’ analyses of unpublished data provided by the Illinois Department of Child and the 1990Ð92 average and the 1997Ð99 Family Services and the Pennsylvania Office of Children, Youth, and Families. The comparison average. is between the 1990Ð92 average and the 1997Ð99 average.

8 Welchans, 2000). The decline in intimate partner assault is particularly noteworthy Figure 12: Trends in Sexual Against Juveniles, 1993Ð2000 because of its connections to sexual abuse. Child sexual abuse is thought to be more 6 common in families where there is intimate partner violence (Rumm et al., 2000). Like 5 child sexual abuse, intimate partner vio- 4 lence has in recent years been the sub- 56% decline ject of substantial publicity, increasingly 3 aggressive efforts at case detection, up- 72% decline graded enforcement activity, and 2 stiffened legal sanctions. 1 Another social problem that has declined markedly over the same period is the Rate per 1,000 Juveniles 0 number of out-of-wedlock teenage preg- 1993 1994 1995 1996 1997 1998 1999 2000 nancies. The rate of live births to teenage mothers dropped 28 percent from a high Ye a r of 39 live births per 1,000 females ages 15–17 in 1991 to 28 per 1,000 in 2000 All offenders Known offenders Unknown offenders (Moore et al., 2001). Teen has often been an outcome of sexual abuse, Note: Juvenile victims are ages 12Ð17. Known offenders are family members or acquaintances, both in the form of conceptions resulting and unknown offenders are or unidentified. from abuse (Boyer and Fine, 1992) and in Source: NCVS, each year, for number of incidents; authors’ analyses. conceptions resulting from the sexualized behaviors that victims sometime manifest in the wake of abuse (Butler and Burton, 1990). The number of teens who reported being currently sexually active or ever Figure 13: Juvenile Sexual Abuse Trends in Minnesota, 1998Ð2001 having also fell during the 1990s (Terry and Manlove, 2000). 8 Yet another related indicator that experi- 7 22% decline enced a decline is the frequency of chil- dren running away. Police reported 28 6 percent fewer incidents of taking children 5 into custody for running away in 1999 than in 1995 (Federal Bureau of Investiga- 4 22% decline tion, 2001). A comparison of data from the 3 Second National Incidence Studies of Miss- Reporting Abuse rcentage of Juveniles ing, Abducted, , and Thrownaway

Pe 2 Children (NISMART–2) with data from 1989 1992 1995 1998 2001 NISMART–1 also indicates a decline in the runaway problem between 1988 and 1999 Survey Year (Hammer, Finkelhor, and Sedlak, 2002). Running away is frequently cited as a strategy used by children who are Nonfamily adult offender Family offender being sexually abused at home, and it is a Note: Respondents are 6th, 9th, and 12th grade students enrolled in public schools in selected behavior that also increases a youth’s risk Minnesota school districts. for sexual assault and exploitation (Famu- Source: Unpublished data from the Minnesota School Survey. laro et al., 1990). The decline in running away is consistent with a decline in sexual abuse. Also consistent with a decline in sexual abuse is a drop in teen suicide. Interagency Forum on Child and Family the motivations (such as discouragement Vital statistics show a modest 18-percent Statistics, 2000). This decline was accom- and ) behind offending. reduction in suicide among 15- to 19-year- panied by a decrease in Divorce rates also declined over the same olds between 1990 and 1999 (Annie E. and a reduction of families on welfare. time period as the social indicators de- Casey Foundation, 2002). Poverty, unemployment, and welfare have generally been viewed more as risk factors scribed above. Divorce rates fell from 4.8 In the same period, the percentage of for forms of child maltreatment other than per 1,000 individuals in 1992 to 4.1 per children living in poverty dropped 27 sexual abuse. It is possible, however, that 1,000 in 2000, a decline of 15 percent (U.S. percent, from 22 percent of children increased employment, particularly among Department of Health and Human Services, younger than 18 in 1992 to 16 percent of potential offenders, may have reduced 2001). Divorce is generally considered both children younger than 18 in 1999 (Federal both opportunities to molest and some of a for and a result of sexual abuse.

9 Greater decline in the most readily pre- Conclusion lack evidence. Indeed, some states clearly ventable cases. Intensive public aware- have made statistical and administrative No solid and convincing explanation exists ness about child sexual abuse, combined changes that have contributed to the for why sexual abuse cases declined in the with aggressive efforts at case finding, decline. There is evidence both that alle- 1990s, although it is important to try to find prosecution, and incarceration, may have gations involving very young children out why a decline occurred. The answer, had some deterrent effect on potential have declined more, perhaps because if it can be determined, is not likely to be offenders. They may be increasingly inhib- such cases have less credibility, and that a simple one. In all likelihood, multiple fac- ited by fears of detection and prosecution. cases involving young perpetrators may tors were involved in the trend. Based on generally has its greatest have declined because they are seen the strength of current evidence, one of effect on those offenders with the most as outside the purview of the child pro- those factors was probably a true decline potential for self-control, the biggest stake tection system. Evidence from at least one in the occurrence of sexual abuse. Changes in conformity, and the fewest other pres- state is consistent with the possibility that in the practices of professionals who re- sures to deviate. Based on this logic, a some of the decline in substantiated cases port suspected abuse and of the child pro- true decline in sexual abuse would be of sexual abuse may be due to a backlash tective system probably also have played expected to occur differentially among against those who report it. a part, but how large a part is difficult to biological fathers in intact families. They ascertain. tend to be the least compulsive offenders Taken together, however, the evidence for these other explanations seems to exist with a lesser tendency to recidivate, the The evidence for some true decline in inci- only in some places or to explain only a most responsiveness to treatment, and dents of sexual abuse comes from several small portion of the decline in substan- a considerable stake in conformity directions. One is the decline in self-report tiated cases. The decline has been so (Hanson, 2001). measures of sexual assault and sexual widespread geographically and has oc- abuse. The NCVS and the Minnesota Stu- State data confirm a particularly large curred across so many categories of chil- dent Survey are both crucial indicators decline in sexual abuse by biological dren, offenders, types of abuse, and types that are independent of the filtering or fathers in intact families. Figures 3 and 8 of evidence that a true decline can be policies of social agencies. Although valid- (see pages 4 and 7, respectively) show a considered as at least one part of the ity problems are always present with the decrease in sexual abuse by parental fig- overall picture. self-reporting of sensitive information, ures and in families with two biological there are no strong reasons to think that or adoptive parents present. This trend Many observers of the decline in the num- candor about sexual abuse has declined. could be interpreted as consistent with ber of substantiated sexual abuse cases, including state officials, have seemed an argument that deterrence has played Another strong piece of evidence for a resistant to the possibility that the num- arole in a true decline. true decline is the improvement in many bers represent a true decline, preferring other indicators of crime, sexual behavior, Increase in the incarceration of offend- almost any other explanation as an alter- and family problems over the same period ers. Among explanations for the general native. This attitude may stem from a con- of time. The decline in these areas sug- decline in crime during the 1990s, the cern that if people believe sexual abuse is gests general movement toward improve- large increase in the incarceration of of- waning, their vigilance and concern about ment in the well-being of children. An fenders has received the most extensive the problem and willingness to support actual decline in the number of sexual empirical support (Blumstein, 2000; Conk- funding will disappear. Increasing num- abuse cases seems more plausible in the lin, 2003). Although detailed data are bers of cases were part of what mobilized context of such a trend than it would if insufficient to conduct a careful analysis people and resources during the 1980s, so the other factors had not improved. of the possible impact of incarceration declining numbers of cases might have on sexual abuse, the overall pattern is More has been focused on child the opposite effect. certainly consistent with the idea that sexual abuse during the past two decades Although social problems go through a increased incarceration played a part in a than on any other form of child maltreat- well-recognized issue/attention cycle and true decline. Surveys of state correctional ment. It should not be surprising that its some changes have occurred in the media facilities suggest that between 1991 and decline would come before and be greater attitude toward sexual abuse, there are rea- 1997, the number of persons incarcerated than that of other forms of maltreatment. sons to doubt that a true decline in inci- in state correctional facilities for sex Prevention and intervention efforts have dence of the current magnitude could, if crimes against children increased 39 per- included school-based prevention educa- recognized, result in a massive desertion of cent, from 43,500 to 60,700 (Finkelhor and tion, treatment programs for juvenile and interest or funding. For one, the public and Ormrod, 2001), after more than doubling adult offenders, and greatly enhanced professional interest in the issue of sexual from 19,900 in 1986. This does not include resources for criminal justice investigation abuse has roots that go far beyond the mat- the many sexual abuse offenders who re- and prosecution. It is reasonable to think ter of whether it involves 50,000 or 150,000 ceive sanctions that do not involve incar- that, given the scale of these efforts, they cases per year, and relate to the now well- ceration for a year or more. Compared have had some success in preventing or established role that it plays in discussions with people who commit other forms of intervening in sexual abuse. child maltreatment, people who commit of family problems, gender relations, sexu- sexual abuse are much more likely to lose The relatively inconsistent evidence for ality, and mental health. The high-profile access to their victims (as a result of court other explanations of the decline in the public and professional role this problem order or divorce) or to lose their liberty number of sexual abuse cases also sup- has achieved in recent years will not easily altogether. The incapacitation of offenders ports the possibility of a true decline in change. Second, the other social problems alone should be expected to have some sexual abuse. As discussed earlier in this discussed above that also have experi- effect on the number of new cases. Bulletin, the other explanations do not enced recent declines do not appear to

10 have suffered any social policy desertion It would also greatly help the analysis of Boyer, D., and Fine, D. 1992. Sexual abuse as a result. , crime, and teen preg- the current decline and future trends if in adolescent pregnancy and maltreat- nancy are all still issues of ongoing serious data systems relating to relevant factors ment. Perspectives 24(1). policy attention, despite their declines, be- would be expanded, enhanced, and im- Butler, J.R., and Burton, L.M. 1990. Re- cause they remain serious problems even proved (Finkelhor and Wells, 2003). Cur- thinking teenage childbearing: Is sexual at reduced levels. The declines may, in rently, data on sex crimes against children abuse a missing link? Family Relations fact, have spurred policy interest because are artificially divided between the child 39:73–80. problems that fester for a long time with- protective system and the law enforcement out improvement in spite of considerable system in a way that prohibits a compre- Centers for Disease Control and Preven- policy attention become frustrating. Policy- hensive assessment of trends in the whole tion. 2000. Tracking the Hidden Epidemics: makers and the public can become discour- problem. Data from state child protection Trends in STDs in the United States. Atlanta, aged and decide that such problems are systems are not gathered in ways that are GA: Centers for Disease Control and beyond immediate solution. Signs of suc- comparable across ; therefore, Prevention. cess from social initiatives can provide the comparisons of the effects of different pol- public and policymakers with energy and icy environments are difficult. In the justice Conklin, J.E. 2003. Why Crime Rates Fell. justification for expanded efforts to rein- area, systematic information is not readily Boston, MA: Allyn and Bacon. force what appears to be working. Of available on the demographics of persons Famularo, R., Kinscherff, R., Fenton, T., course, the factors influencing public inter- prosecuted, convicted, incarcerated, or and Bolduc, S.M. 1990. Child maltreatment est and policymaking are complex, but treated for sex crimes against children. histories among runaway and delinquent there is no strong reason to believe that children. Clinical 29(12):713–718. evidence of a true decline in sexual abuse In addition, an understanding of the rea- by itself will have negative effects on the sons for the decline has been greatly ham- Federal Bureau of Investigation. 2001. policy environment around the problem. pered by the failure of communities to Crime in the United States, 1999: Uniform evaluate their varied prevention and inter- Crime Reports. Washington, DC: U.S. Because social policy benefits from under- vention efforts. More effort should be Department of Justice. standing the factors that result in success, made prospectively to observe trends and the hypothesis that sexual abuse has de- outcomes as communities implement vari- Federal Interagency Forum on Child and clined should be accepted, and identifying ous prosecution, treatment, community, Family Statistics. 2000. America’s Children: the reasons why it has declined should be and school-based educational efforts. In Key National Indicators of Well-Being, 2000. a priority. It is extremely important that this way, a better inventory of the more Washington, DC: U.S. Government Printing lessons be drawn from a change of this and less successful strategies could be Office. magnitude in a social problem that has been tracked in conjunction with the relative Finkelhor, D., and Ormrod, R.K. 2001. considered so widespread and corrosive decline in different locales. Offenders Incarcerated for Crimes Against to the well-being of children, families, and Researchers may not be able to fully answer Juveniles. Bulletin. Washington, DC: U.S. communities. Several initiatives might be Department of Justice, Office of Justice considered to deepen our understanding. the question of why this most recent de- cline has occurred; however, it is impor- Programs, Office of Juvenile Justice and First, more intensive studies need to be tant to be better prepared to understand Delinquency Prevention. undertaken in individual localities where a the sources of any continuing or future Finkelhor, D., and Wells, M. 2003. Improv- full inventory of explanations could be con- declines. To what extent do prevention ing national data systems about juvenile sidered, with both quantitative and quali- education, increased public awareness, victimization. Child Abuse & Neglect tative evidence available. In individual lo- greater prosecution, and incarceration 27(1):77–102. calities, it may be easier to observe how roles? Answering such questions can policy and programmatic changes, includ- help policymakers formulate policies that Hammer, H., Finkelhor, D., and Sedlak, A.J. ing prosecution initiatives, treatment re- will extend and accelerate the decline in 2002. Runaway/Thrownaway Children: Na- sources, and educational programs, may sexual abuse and, perhaps, in other forms tional Estimates and Characteristics. Bul- have been sequenced with the onset or of child maltreatment. letin. Washington, DC: U.S. Department of acceleration of a decline in sexual abuse. Justice, Office of Justice Programs, Office In addition, localities with different trend of Juvenile Justice and Delinquency patterns (steady declines, increases, no References Prevention. change, and fluctuating patterns) should Annie E. Casey Foundation. 2002. Kids Hanson, R.K. 2001. Age and Sexual Recidi- be compared with one another, and it Count 2002. Retrieved March 11, 2003, vism: A Comparison of Rapists and Child might be useful if such localities were in from www.aecf.org/kidscount/kc2002/ Molesters. Ottawa, Ontario, : Solic- the same state and were comparable in pdfs/summary.pdf. itor General of Canada. other ways. Some local studies might cen- Blumstein, A. 2000. Disaggregating the ter around the case records of investiga- Harrison, P.A., Fulkerson, J.A., and Beebe, violence trends. In The Crime Drop in tive agencies that have maintained stable T.J. 1997. Multiple substance use among America, edited by A. Blumstein and J. policies, catchment areas, and detailed adolescent physical and sexual abuse vic- Wallman. Cambridge, England: Cambridge recordkeeping practices over a long peri- tims. Child Abuse & Neglect 21(6):529–539. od, from which it might be ascertained University Press, pp. 13–44. more accurately how case characteristics Hechler, D. 1988. The Battle and the Back- have changed over time. lash: The Child Sexual Abuse . Lexing- ton, MA: D.C. Heath.

11 U.S. Department of Justice PRESORTED STANDARD Office of Justice Programs POSTAGE & FEES PAID Office of Juvenile Justice and Delinquency Prevention DOJ/OJJDP *NCJ~199298* PERMIT NO. GÐ91 Washington, DC 20531

Official Business Penalty for Private Use $300

Bulletin NCJ 199298

Jones, L., and Finkelhor, D. 2001. The Rennison, C.M., and Welchans, S. 2000. The Office of Juvenile Justice and Delinquency Decline in Child Sexual Abuse Cases. Intimate Partner Violence. Special Report. Prevention is a component of the Office of Bulletin. Washington, DC: U.S. Department Washington, DC: U.S. Department of Jus- Justice Programs, which also includes the of Justice, Office of Justice Programs, tice, Office of Justice Programs, Bureau Bureau of Justice Assistance, the Bureau of Office of Juvenile Justice and Delinquency of Justice Statistics. Justice Statistics, the National Institute of Prevention. Rumm, P.D., Cummings, P., Kraus, M.R., Justice, and the Office for Victims of Crime. Jones, L.M., Finkelhor, D., and Kopiec, K. Bell, M.A., and Rivara, F.P. 2000. Identified 2001. Why is sexual abuse declining? A spouse abuse as a risk factor for child survey of state child protection adminis- abuse. Child Abuse & Neglect 24(11): trators. Child Abuse & Neglect 25(9): 1375–1381. Acknowledgments 1139–1158. Terry, E., and Manlove, J. 2000. Trends This Bulletin was written by David Minnesota Department of Children, Fami- in Sexual Activity and Contraceptive Use Finkelhor, Ph.D., Professor of Soci- & Learning, Minnesota Department of Among Teens. Research Brief. Washington, ology, and Director, Crimes against Human Services. 2001. Minnesota Student DC: Child Trends. Children Research Center, Universi- Survey: Key Trends Through 2001. Roseville, ty of New Hampshire; and Lisa M. U.S. Department of Health and Human MN: Minnesota Department of Children, Jones, Ph.D., Research Assistant Pro- Services. 2001. National Vital Statistics Families & Learning. fessor of , Crimes against Report: Annual Summary of Births, Mar- Children Research Center. For more Moore, K.A., Manlove, J., Terry-Humen, E., riages, and Deaths: United States, information on this research, contact Williams, S., Papillo, A.R., and Scarpa, J. 1992 49(6). Washington, DC: U.S. Govern- the authors at 603Ð862Ð1888, david. 2001. CTS Facts at a Glance. Washington, ment Printing Office. [email protected], or lmjones@unh. DC: Child Trends. edu, or visit the Center’s Web site at U.S. Department of Health and Human www.unh.edu/ccrc. Myers, J.B. 1994. The Backlash: Child Pro- Services, Office of Child Abuse and tection Under Fire. Thousand Oaks, CA: Neglect. 1992–2002. Child Maltreatment The authors would like to thank staff Sage Publications, Inc. 1990–2000: Reports from the States to the from the following departments for National Child Abuse and Neglect Data their permission and assistance in Current Peddle, N., and Wang, C.T. 2001. System. Washington, DC: U.S. Government obtaining data used in this report: Trends in Child Abuse Prevention, Report- the Pennsylvania Office of Children, Printing Office. ing, and Fatalities: The 1999 Fifty State Youth, and Families; the Illinois Survey. Working Paper. Chicago, IL: Department of Child and Family Prevent Child Abuse America. This Bulletin was prepared under grant number Services; the Oregon Department 98–JN–PX–0012 from the Office of Juvenile of Human Services; the Minnesota Rennison, C.M. 2001. Criminal Victimiza- Justice and Delinquency Prevention, U.S. Department of Human Services, tion 2000: Changes 1999–2000 With Trends Department of Justice. Children’s Research, Planning and 1993–2000. Bulletin. Washington, DC: U.S. Evaluation Department; and the Department of Justice, Office of Justice Points of view or opinions expressed in this Minnesota Department of Children, Programs, Bureau of Justice Statistics. document are those of the authors and do not Families & Learning. necessarily represent the official position or policies of OJJDP or the U.S. Department of Justice.