DUE 8/27

Childhood Victimization and Subsequent Risk for Promiscuity, Prostitution, and Teenage Pregnancy: A Prospective Study

A B S T R A C T Cathy Spatz Widom, PhD, and Joseph B. Kuhns, MA

Objectives. This study examined the extent to which being abused and/or neglected in childhood in- creases a person s risk lor promiscu- ity, prosiitution, and teenage preg- nancy.

Methods. A prospective cohorts design was used to match, on the basis of age, race, sex, and social class, cases of abused and/or ne- glected children from 1%7 lo 1971 with imnabused and nonneglected children; subjects were followed into young adulthiMid. From 1989 to 1995, 11% subjects (67f) abused and/or neglected and 520 control subjects) were located and interviewed.

Result,s. Early childhiH)d abuse and/or neglect was a significant pre- dictor of prostitution for females (odds ratio 1OR| = 2,%). For fe- males, sexual abuse (OR = 2,54) and neglect (OR = 2.58) were associated with prostitution, whereas physical abuse was (inly marginally associ- ated. ChildluHid abuse and neglect were not asstKiated with increased risk for promiscuity or teenage preg- nancy.

Conclusions. These findings strongly support a relationship be- tween childhood victimization and subsequent prostitution. The pre- sumed causal sequence between childhood victimization and teenage pregnancy tnay need to be reevalu- ated. {Am J Public Health. 1996:86: 1607-1612)

Introduction

Clinical and behavioral reports and research studies have linked childhood victimization to a variety of negative health and behavioral consequences. Pros- titution.'"'^ promiscuity,"""-'' and teenage pregnancy'^"'"" have been included among these sequelae. From a puhlie health perspective, these consequences of abu- sive and neglectful earlychildhmnl experi- ences are particularly important in terms of potential sexually transmitted diseases, unwanted and early pregnaneies, and the potential to facilitate a youth's transition into prostitution. In addition, any relation- ship between childhood victimization and teenage pregnancy is important to un- cover because of the link between prema- ture parenting and inadequate child- rearing practices, factors that perpetuate ihe cycle of abuse from one generation to the next.

Existing research on the connections between childhood victimization and pro- miscuity, prostitution, and teenage preg- nancy varies in terms of study designs and lindings. One of the major limitations of this body of research is its reliance on correlational (cross-sectional) designs, with data collected at one point in time. Correlational studies do not permit exami- nation of causal sequences. For example, one concern in interpreting this literature is that a teenager's pregnancy tnay have resulted from forced sexual intercourse or from promiscuity or irresponsible contra- ceptive protection following a history of sexual abuse.^^ Another possibility is that the pregnancy could be planned as an escape from an abusive home environ- ment.-' While the majority of this re- search is based on the belief that child- hood victimization leads to certain outcomes, it is possible that some of these problem behaviors precede abuse or

neglect. Finally, reliance on retrospective accounts of childhood victimization leaves the data open to a host of potential biases.'''"'**

With some exceptions, designs lack appropriate comparison or control groups. Since child abuse is often associated with low-income families,'"*' comparison or control groups are important. Low- income families often have many prob- lems, child abuse being only one of them. Because other factors, such as poverty, unemployment, parental alcoholism, drug problems, or otherwise inadequate soeial and family functioning, arc often present in such muitiproblcm homes, controi groups matched on socioecontimic status and other relevant variables become vital components of this researeh.'" The child sexual abuse literature has been fairly vague in separatingcliccts directly related to sexual abuse from effects possibly due to preexisting psychopathology in the child, family dysfunction, or the stress associated with disclosure.''- Such con- founding of eorrelated characteristics needs to be avoided so that the unique contribution of childhood sexual abuse, physical abtise. or neglect can be disen- tangled from other background character- istics, ln sum, while there is correlational evidence that chiidhood victimization may be asstK'iated with increased risk of prostitution, promiscuity, and teenage pregnancy, current knowledge of the extent and nature of these relationships is

The authors are with the Schtxil of Criminal Justice. The University at Albany. Albany. NY.

Requests for reprints should be sent to Cathy Spatz Widom, PhD. ThL* University at Albany. Schtml of Criminal Justice. 1.15 West- ern Ave. Albany. NY 12222.

This paper was accepted July 3, 1996. Note. The views expressed liere are the

authors' and do not necessarily represent those of the US Department nl Justice.

November 1996. Vol. 86, No. 11 American Journal of Public I lealth 1607

Widuni and kuhns

ambiguous, at best, in the absence of longitudinal research.

This paper examines the relationship between early childhood victimization and subsequent promiseuity, prostitution. and teenage pregnancy. The cohort de- sign and analytie teehniques used permit- ted examination of the nature of these relationships longitudinally while control- ling for relevant demographic variables. On the basis of previous literature, the overall hypothesis tested was that there would be a direct relationship between early childhiwd victimization and subse- quent promiscuity, prostitution, and teen- age pregnancy. It was also hypothesized that victims of childhood sexual abuse would be at a particularly increased risk for promiscuity, prostitution, and teenage pregnancy.

Methods

In the current study, abused and neglected children were matched with nonabuscd and nonncgleeted children and followed prospectively into young adulthood.'"^' (Characteristics of the pro- speetive design ineluded (1) an unambigu- ous operationalizalion of abuse and ne- glect; (2) separate abused and neglected groups; (3) a large sample; (4) a eontrol group matehed as closely as possible for age, sex, race, and approximate family social class background; and (5) assess- ment of the long-term consequences of abuse and neglect beyond adoleseenee and into adulth(X)d (for eomplete details of the study design and subjeet selection criteria, see Widom""),

It was important to match for social elass in this study because it is theoreti- eally plausible that any relation between child abuse or neglect and the outcomes under investigation would be confounded or explained by stKial class differences. It is difficult to niateh exactly for social class because higher ineome families may live in lower social class neighborh(x>ds, and vice versa. However, the matching priKe- dure used here was based on a broad definition of soeial class that ineluded neighborhoods in which children were reared, schools they attended, and hospi- tals in whieh they were bom. Busing was not operational at this time, and students in elementary schools in the study county were from small, socioeeonomically ho- mogenous neighborhoods.

Subjects

The abused and neglected group comprised individuals with substantiated

cases of childhood physieal and sexual abuse and/or neglect processed during the years 1967 through 1971 in the eounty juvenile or adult criminal eourt (situated in a metropolitan area in the Midwest). These cases of early abuse and/or neglect were restricted to ehildren who were 11 years of age or younger at the time of the incident (n = 908).

Physical abuse eases included inju- ries such as bruises, welts, burns, abra- sions, laeerations, wounds, cuts, bone and skull fraetures, and other evidence of physical injury. Sexual abuse cases varied from relatively nonspecific charges of "assault and battery with intent to gratify sexual desires" to more specific ones including "fondling or touching in an obseene manner," sodomy, incest, and so forth. Neglect cases reflected a judgment that the parents' deficiencies in child care were beyond those found acceptable by community and professional standards at the time. These cases represented ex- treme failure to provide adequate fot>d, clothing, shelter, and medical attention to ehildren.

A critical component of this research involved the identification of a matehed eontrol group. Children who were under school age at the time of the abuse and/or negleet ineident were individually matched with children of the same sex, raee, date of birth (±1 week), and hospital of birth through the use of eounty birth reeord information. For children of sehool age at the time, elementary sehool reeords for the same time period were used to find matehes with children of the same sex, race, date of birth (±6 months), and same elass in the same elementary sehool during the years 1967 through 1971. Overall, there were matehes for 74% of the abused and neglected ehildren.

Nonmatehes occurred for a number of reasons. In terms of birth records, nonmatehes occurred in situations in whieh the abused or negleeted ehild was bom outside the county or state or date-of-birth information was missing. In terms of school records, nonmatehes oceurred because class registers were unavailable as a result of the elosing of the elementary sehool over the past 20 years or lack of adequate identifying informa- tion for the abused and neglected chil- dren.

Initially, offieiai eriminal histories of the abused and neglected children were examined and compared with those of the eontrol group."" A second phase of the researeh involved the locating and inter- viewing of these abused and/or neglected

individuals (20 years after their childhcjod victimization) and control subjects to document the long-term consequences of ehildhood victimization across a number of domains of functioning (cognitive and intelleetual, emotional, psychiatric, sixrial and interpersonal, oecupational, and gen- eral health). Two-hour follow-up inter- views, conducted between 1989 and 1995, eonsisted of a series of structured and semistructured questions and rating scales, measures of 10 and reading ability, and a psychiatric assessment. The interviewers were unaware of the purpose of the siudy and of the inclusion of an abused and neglected group. Similarly, the subjects were unaware of the purpose of the study. That is, subjects were lold that they had been asked to participate as part of a large group of individuals who grew up in that area during the late !96()sandearly I97()s. Subjeets who participated signed a eon- sent form acknowledging that they were participating voluntarily.

The findings described here are based on completed interviews with 1196 individuals. Of the original sample of 1575, 1291 subjects (82%) were loeated and 1196 were interviewed (76% overall, representing 74% of the abused and neglected group and 78% of the aintrol group). Of the 95 people not interviewed, 39 were deceased, 9 were incapable of being interviewed, and 49 refused to participate (a refusal rate of 3%).

Comparison of the current follow-up sample with the original sample indicated no significant differences in terms of percentage male. White, or abused and/or neglected; poverty in childhtwd census tract; or mean current age (abuse/negleet group = 28.6 years, control group = 28.9 years). The interviewed group (follow-up sample) was significantly more likely lo have bad an offteial criminal history than the original sample of 1575 (42% vs 36%). However, this is not surprising sinee people wilh a criminal history arc gener- ally easier to find, in part because they have more "institutional footprints" to assist in loeating them.

Variables

Teenage pregnaney was defined as having had a child before 18 years of age. This was based on a calculation of the age of the oldest child in comparison with the age of the person at the time of the interview. Subjects who indicated that they were less than 18 years old at the birth of their first child were eoded as having had a teenage pregnaney (n = 159). Subjects were defined as promiscuous if

1608 American Journal of Public Health November 19%. Vol. 86. No. 11

Child Abuse and Sexual Behaviors

they reported having had "sex with 10 or more people within any single year" (n = 207). Prostitution was defined as a positive response to a question about whclher the person had "ever been paid for having sex with someone" (n ^ 101). These ouleome variables were dichoto- nious (coded 1 orO).

Independent variables ineludedehild- hood vietiniization as well as controi variables of age, sex, and race. Group was a dichotomous variable based on official reports of early ehildh(K>d abuse and/or neglecl (0 = control. I = abuse/neglect). In analyses focusing on consequences associated with dilTerent types of abuse or neglect, three categories (separate dummy variables) based on oflicial reports of childhood victimization were used: atiy physieal abuse, any sexual abuse, and any negleet. It should be noted that respon- dents may have experienced more than one type of abuse or negleet.

Multivariate analyses controlled for sex, race/ethnicity, age, and approximate family social class. Raee/ethnieity was based on respondents' self-identification. White non-Hispanic was coded I; all other raee/ethnieity categories (Hispanic and non-Hispanie Black, Hispanic White, American Indian, and Pacific Islander) were coded (I. Age was an interval-level variable indicating the person's age at the lime ot the interview. The mean age of the sample at the time of the interview was 2y.2 years (SD = 3.S). As an indicator of the fatnily's approximate social class dur- ing the person's childhood, a variable labeled "welfare as a child" was included. A positive score (1) on this dichotomous variable rellected the respondent's report that hisor her family had received welfare when the respondent was a child,

Initial results are presented in terms of bivariatc statistics, Multivariate analy- ses using logistic regression were con- ducted to determine whether the findings at the bivariate level would be supported in adjusted analyses after demographic characteristics had been controlled. Logis- tic regression equations were estimated with age, race, sex, and welfare as a child as eontrol variables predicting the dichoto- mous dependent variables of prostitution, promiscuity, and teenage pregnanc7. A second set of equations was estimated to examine the contributions of different types of abuse or neglect. Analyses were performed for the overall sample and for male and female subjects separately. These analyses were also replicated with matched pairs only and with a version of conditional maximum likelihood logistic

TABLE 1—Prevalence of Promiscuity, Prostitution and Teenage Pregnancy in a Cohort of 676 Young Adults Abused as Children and in 520 Control Subjects

Overall Abuse/neglect group

Neglect Physical abuse Sexual abuse

Control group

Female subjects Abuse/neglect group

Neglect Physical abuse Sexuai abuse

Control group

Male subjects Abuse/neglect group

Neglect Physical abuse Sexual abuse

Control group

No.

676 543 110 96

520

338 257 47 76

244

338 286 63 20

276

Promiscuity, %

1779 18.92 19.63 13 68 16.99

7.14 7,84

12.77 6.56 5.74

28.53 28.67 25.00 42.11 27.01

Prostitution,

10.73** 11.13*** 12,04* 9,38 5.60

8.93'* 9.02**

12.77** 10.53** 2.87

12.54 13,03 11,48 5.00 8.03

Teenage % Pregnancy, %

14.35 15,29 11,82 12.50 11.92

23.96 26.46 25.53 15.79 20.90

4.73 5.24 1.59 0.00 3,99

Note. All statistical comparisons (chi-square tests) were made with the conlrol group. Respondents could be coded 1 for more than one type of abuse (thus. the sum of the abuse/neglect group si2GS does not equal the total number of abused/neglected subjects).

*P < .05;**P < . 0 1 ; " ' P < .001 (two tailed tests).

regression. However, since all subjects, not only matched pairs, were used in the analysis, conditional analyses proved infca- sibie. The results of these analyses were essentially the same as those presented here.

Results

The first hypothesis examined whether experiencing childh(K)d victimiza- tion was associated with increased risk lor promiscuity, prostitution, or teenage preg- naney. Table I presents the prevalence of promiscuity, prostitution, and teenage pregnancy by group, gender, and type of abuse or neglect. The most striking iind- ingwas the lackof differenee between the abused/neglected group and the control group {overall and for male and female subjects separately) in terms of promiscu- ity and teenage pregnancy (see Table I).

Only the likelihw)d of engaging in prostitution differentiated childhood abuse and neglect victims from individuals who did not have ofiieial records of abuse or neglect, and a significant increase in risk for prostitution was eonfined to female subjeets. The ditferenee in prevalence of prostitution for abused and negieeted male subjects vs male control subjects approached, but did not reach, the conven- tional level of significance. Overall, about

9% of the abused and neglected female subjects, as compared with ^% of the female eontrol subjects, reported having engaged in prostitution. Physieally abused female subjeets had the highest rates (12.8%) of prostitution, followed by sexu- ally abused (10.5^^ ) and neglected (9.0%) female subjects.

Multivariate results are presented separately for promiscuity (Table 2), prostitution (Table 3), and teenage preg- nancy (Table 4), Group (abuse/neglect vs control) was not a significant predictor of promiscuity for the sample overall or for male and female subjects separately (Table 2). Furthermore, none of the specifie types of abuse or neglect predicted promis- cuity (Table 2).

Group (i.e., childh(X)d abuse or ne- glect) remained a significant predictor of prostitution, after control for demo- graphic characteristics and welfare as a ehild, for the sample overall and for female subjects (Table 3). For male subjects, the effect of group did not reach the customary level of signilicance {P = .13), suggesting that the effect was confined to female subjects. The interac- tion between gender and group was not signifieant {P < .14). ChildhtHxJ sexual abuse was a signilicant predictor of prosti- tution in female subjects (Table 3), Ne- gleet also predieted prostitution for the

November 19%, Vol. 86. No. 11 Americim Joumul of Public Health 16(W

Widom and Kuhns

TABLE 2—Logistic Regressions Predicting Promiscuity in Young Adults Abused as Chiidren

Overall (n = 1187)

Female Subjects (n = 580)

Male Subjects (n = 607)

OR 95% Ci OR 95% Ci OR 95% Ci

Abuse/neglect

Type of abuse Sexual abuse Physical abuse Neglect

1.06 0.77,1.48 1.07 0.53.2.16 1.06 0.73,1.54

1-57 1.10 1.14

0.81,3.06 0.65, 1.86 0.83, 1.59

1.17 1.89 1.22

0.42. 3.22 0.73, 4.87 0.61,2.46

2.37 0.91,6.16 0.90 0.48, 1.67 1.10 0.76.1.60

Note. Sample sizes given are both case and control subjects. All statistical comparisons were made wrth the controi group. Age, race, gender, and an indicator ot socioeconomic status (receiving welfare as a child) were controlled in the regression equations. OR = odds ratio; Cl = confidence intervai.

TABLE 3—Logistic Regressions Predicting Prostitution in Young Aduits Abused as Chiidren

Overall (n = 1189)

Female Subjects (n = 580)

Male Subjects (n = 609)

OR 95% Ci OR 95% Cl OR 95% Cl

Abuse/neglect

Type of abuse Sexual abuse Physical abuse Neglect

1.88'

1.73 1.79 1.78*

1.19.2.99 2.96* 1.26.6.96 1.54 0.88,2.71

0.80, 3 78 0-94, 3.40 1.15,2.77

2.54* 2.63 2.58'

1.02,6-32 0.99, 6.96 1.20.5.55

0.63 1.35 1.48

0.08, 5.00 0.58.3.18 0.86, 2.55

Note. Sample sizes given are both case and controi subjects. All statisticai comparisons were made with the cortroi group. Age, race, gender, and an indicator of socioeconomic status (receiving weifare as a child) were controlled in the regression equations. OH = odds ratio; Cl = confidencs interval.

•P < -05; **P < .01 (two-tailed tests).

TABLE A—Logistic Regressions Predicting Teenage Pregnancy in Young Aduits Abused as Chiidren

Overall (n = 1196)

Femaie Subjects (n = 582)

Male Subjects (n = 614)

OR 95% Cl OR 95% Ci OR 95% Cl

Abuse/negieet

Type of abuse Sexual abuse Physical abuse Negieet

1.10 0.76,1.59 1-11 0.73,1.67 1.07 0.47,2.40

0.62 1.05 1.30

0.32. 1.20 0.55, 1.98 0-90, 1.87

0.65 1.31 1.31

0.33. 1.28 0.65, 2.64 0-87, 1 98

Note. Sample sizes given are both case and control subjects. All statistical comparisons were made with the control group. Age, race, gender, and an indicator of socioeconomic status (receiving welfare as a child) were controlled in the regression equations. OR = odds ratio; Ci = confidence interval.

sample overall and for female subjects, whereas the effect for male subjects did not reach the customary level of signifi- cance {P = .16). It should be noted that.

with no control for demographie charaeter- istics or approximate family social class (Table I), physically abused female sub- jects had significantly higher rates of

prostitution. However, after these factors were controlled for (Table 3). the physical abuse effect became more borderline and not significant.

Table 4 presents the results of logis- tic regression equations predicting teen- age pregnancy by group (abuse/neglect vs control) and by type of abuse or neglect. Surprisingly, neither group nor type of abuse or neglect was a significant predic- tor of teenage pregnancy for either the overall sample or the females. This find- ing contrasts dramatically with previous literature, although differences between the present design and others may help explain the discrepancy.

Discussion

Despite widespread belief that ehild- htxid victimization is associated with promiscuity, prostitution, and teenage pregnancy, our (indings. based on a prospective eohort design, indicate that childhood victimization is not a signifi- ieant risk factor for promiscuity or teen- age pregnancy. That is. we found no significant relationship between early chi!dh(Mid abuse/neglect and promiscuity and teenage pregnancy, either in bivariate or in muitivariale analyses Ihat controlled for age. race, sex, and welfare status as a child. Thus, our results do not conform to general expeetations or lo previous re- ports about these two outcomes. How- ever, this study differs from most past research in the prospective nature of the design and the inclusion of a control group. The prospective nature of this study allows some issues of causality to be examined and disentangles the effects of childhtxtd victimization from other poten- tial eonfounding effects.

The outeomes that other researchers have found and attributed to early child- h(K>d victimization experiences (particu- larly sexual abuse) may nol be specifically associated with childhtxjd victimization or sexual abuse in particular. Although a high percentage of young women become pregnant a.s teenagers and many report having experienced earlier sexual abuse, there may be other factors that explain the apparent relationship. For example, Leibowitz et al."" adopted a primarily economic perspective to explain teenage pregnancy, whereas Owksey''^ examined the effects of family background factors, including family structure and size, paren- tal education and employment status, and religious affiliation. Cooksey*' showed that many background factors predicted premarital pregnancy (e.g., there was a

American Journalof Public Health November 1996. Voi. 86, No. 11

Child Abubc and Sexual Behaviors

higher likelihood for Blacks than for Whites or Hispanics and a lower likeli- hood for individuals with an intact family structure and few siblings).

Thus, high rates of teenage preg- nancy may be found in young women who come from economically deprived back- grounds (i.e.. from families who may have multiple problems, both including and excluding chiid abuse or neglect). In previous studies not involving a demo- graphically matched control group, this patteni would not have been evident. Also, because studies often involve preg- nant teenagers who are asked about their earlier backgrounds, abused and ne- gieeted children who did not become pregnant would not be included, thus providing an incomplete picture of the long-term consequences of early child- h(K)d victimization.

On the other hand, we did find that early childhood victimization was assoei- ated with increased risk for prostitution, although this was true only for female subjects. These findings reinforee the earlier literature calling attention to the role of ehildhood victimization in the backgrounds of prostitutes. Early child- hood abuse and negleet appear to place children at increased risk of becoming prostitutes, whieh reinforces the impor- tance of viewing prostitution in a victimiza- tion context.

Despite an emphasis on ehildhmid sexual abuse as an antecedent to prostitu- tion, we found that ehildhood negleet is also associated with inereased risk for prostitution. By definition, childhtx)d ne- glect involves inadequate supervision and inadequate caretaking by a parent or guardian. Bracey^ described juvenile pros- titutes as being neglected at home, quot- ing one young woman who reported that her "parents didn't tell me to get out. but they didn't come looking for me when I did." Negieeted young children, on the streets alone, are at risk of being vietim- ized or enticed into prostitution. For example, runaways may come under the eontrol of pornographers and pimps and bea)me suseeptible to subsequent physi- cal and sexual victimization by pimps and eustomers.''^

We have also found that abused and negieeted ehiidren are at inereased risk, relative to nonabused and nonnegleeted control children, of being picked up by the police for running away** and of reporting having run away overnight. The introduc- tit)n of "running away" into our regression equations to estimate its role as a media- tor between childhtwd victimization and

prostitution reduced the contribution of ehildhood victimization as a predictor of prostitution overall. However, ehildhood victimization remained a significant predic- tor (data not shown).

One issue in considering these prosti- tution findings, however, is the extent to whieh the relationship between early childhood victimization and prostitution is part of a larger syndrome of problem behaviors for whieh abused and neglected children are at risk.'**' Previous research has indicated that abused and negieeted children are at an increased risk of being arrested as a juvenile, as an adult, or for a violent crime."*'' Future analyses will at- tempt to disentangle whether the in- creased risk of prostitution described here is specific (and should be interpreted in the victimization context) or whether it is part of a lifestyle that plaees abused and negieeted children at inereased risk in general for criminal behavior.

Another issue eoncems the role substance abuse plays in relation to prostitution. Among Bagley and Young's^ sampie of pnwtitutes, the primary reason for entrance into prostitution reportedly was the need to suppt-irt a drug habit. Seng's'' sample of prostitution-involved children reported signihcantly higher rates of alcohol and drug abuse than did the sexually abused children who were not involved in prostitution. Other studies have reported correlations between pros- titution and drug use.̂ *^ '̂' Although such an analysis was beyond the seope of the present paper, we plan to examine the relationship between substance use and abuse and prostitution (particularly look- ing at temporal sequences) in future research.

While this study design has clear strengths, certain caveats should also be kept in mind. First, these eases of child- hood abuse and negleet came to the attention of the authorities and were skewed toward the lower end of the soeioeconomic spectrum.̂ "* Members of our matehed control group eame from the same neighborhoods as the abused and negieeted children and were also from lower soeioeconomic classes. This means that the sample is not representative of the population as a whole and that these findings cannot be generalized to all cases of abuse and neglect. For example, middle- and upper-class women who might have e.xperienced childhood sexual abuse were less likely to be included here.

Seeond, the teenage pregnancy mea- sure used here may have represented an

underestimate of the extent of teenage pregnancies experienced. Our measure did not include women who were preg- nant but whose pregnaneies ended in abortions or miscarriages. It is possible that women whose children died might have failed to mention these ehiidren. We hope that other researchers engaged in prospeetive research with different samples from different geographic areas and time periods will examine these issues and attempt to replicate our findings. D

Acknowledgments This ri;soiirch was suppitrlcd hy grants from the Nation:il Institutei>l Menial 1 lealth (Ml 144467) iindthcNalional lnstituteof Justice (S(>-IJ-C'X- 1)0.13 and S')-IJ-CX'(XHI7).

We express great appreciation U) Patricia Glynn for her help with data management and statistical analyses and to Dcbtmih Coppola for her help in preparing the manuscript.

References 1. Allen D. Young male prostitutes: a psycho-

stKial study. Arxh Sex Behav. 1980;9:399- 426.

2. Bagley R. Sexual Offen.\es against Children. New York. NY: Putnam. 1984.

3. Bagley C Young L. Juvenile prostitution and child sexual abuse: a controlled study. Can J Community Ment Health. iy87;6:5- 26,

4. Bracey D. The juvenile prostitute: victim andofl'cnder. V1ctimolog\'. I%3:S:ISI.

5. Burgess AW. Martman CR. MeCorniack A. Abused to abuser: antecedents of socially deviant behaviors. Am J P.'iychiatry.

6. Earls CM. David. Early family and sexual experiences of male and female prosti- tutes. Canada's Ment Health. December 1990:7-11.

7. Greenwald H. The Elegant Prostitute. New York. NY: Ballantine Books; 1970.

S. James J. Meycrding J. Early sexual experi- ence as a iaclor in prostitution. Arch Sex Behav. 1977;7:3i-12.

9. Seng M. Chiid sexual abuse and adolescent prostitution: a comparative analysis./4(/o/es- cence. 1989:24:66S-67.S.

10. Silbert M. Pines A. Sexual child abuse as an antecedent of prostitution. ('hild .Ahu.se Negl. 19KI;5:4l)7^1l.

11. Silhert M. Pines A. Early sexual exploita- tion as an influence in prostitution. Soc Work Res Abstracts. 1983;28:28.S-290,

(2. Simons RL. Whitbeck LB- Sexual abuse as a precursor to proslimiion and victimiza- tion among adolescent and adult homeless women. / fwi /.V.VM<'.V. I Wl:12:.161-37<).

13. Schaffer B. DeBlassie R. Adolescent pros- titution. Adolescence. 1984:19:f,«9-r>9(>.

14. Newton-Ruddy L. Mundelsnian M, Jung- ian feminine psychology and adolescent prostitutes..4(/(>/t'.sr(7icc. l9Sfi:2l:211.

15. Widom CS. Ames MA. Criminal conse- quences of childhtxid sexual victimization. Child Abuse .Negl. IW4:1S:.1(I3-.11S.

If). Conte JR. The efFeci iif sexual abuse on children: a critique and suggestions for

November 19%, Vol. 86, No. 11 American Journal of Public Health 1611

Widom und Kuhns

further research. Victimology. 1985;10:1IO- 130.

17- Browne A, Finkelhor D. Impact of child sexuai abuse: a review of the research. Psychological Bull. 19S6;<J9:66-77.

18. Burgess AW. The Savual Victimization of Adole.scents. Washington. DC: NalJonal Institute of Mental Heiiith. i985.

19. Courtois C A. The incesi experience and its aftermalh. yictimoU)^}'. i979;4:.'^37-347.

20. DeYoung M. The Se.xual Victimization of Children. Jtitft-rson. NC: McFarland Press; 1982.

21. Finkelhor 13. Sexually Victimized Children. New York. NY: Free Pre.ss: iy79.

22. Finkeihor D. Holailing G. Lewis I. Smith C. Sexual aiiuse and iis rehilionship lo laler sexual satisfaction, martiai status, religion, and altitudes. J Intcrpers Viol. 1989:4:379- 399.

23. Herman J. Hirschman K. Father-daughter incest: a feminist theoretical perspective. 7 IVomen Cult. Soc. i977;2:735-756.

24. Lukianowicz N. Incest I: parental incest.Sr J P.mhiatn. 1972; 12t): 130-133.

25. Cavaiola A. Schiff M. Behavioral sequelae of physicai and/or sexuai abuse in adoles- cents. C/i/W..4fcm-£-M'g/. 1988:12:181-188.

26. Laumann EO. Oagnon JH. Michael RT. Michaels S. The Social Organization of Sexuality: Sexual Practices in the United States. Chicago. lil: University of Chicago Press: 1994.

27. Buiier JR. Burton LM. Rethinking teenage childhearing: is sexuai abuse a missing link? Fam Relations. 1490:39:73-80.

28. DeFraneis V. Protecting the Child Victim of Sax Crimes Committed by Adults. Denver, Colo: Americiin I lumane Asstxiiation. 1969.

29. Donaldson PC. Whalen MH. Anastas JW. Teen pregnanc\- and sexual ahuse: explor- ing the connection. .S>;j(V/i Coll Sttid Soc Work. 1989;59:289-3(K).

30. Lantlry E. Bertrand J. Cherry J. Rice J. Teen pregnancy in New Orleans: factors ihal differentiate leens who deliver, abort or successfully contracept.7 Youth Adoles- cence. 1986:1.5:259-274.

31. O'Brien JD. The efTectsof incest on female adolescent behavior. / Am Acad Psycho- anal. 1987:15:83-92.

32. Oz S. Fine M. A comparison of childhood backgrounds of teenage mothers and their non-molher peers: a new formulation. J Adolescence. 1988; 11:25 i -261.

.33. Gershenson HP, Musick JS. Ruch-Ross HS, Magee V. Rubino KK, Rosenberg D. The prevaienee of coercive sexual experi- ence among teenage mothers. J Interpcrs Viol. 1989;4:2()4-2I9.

34- Boyer D. Fine D. Sexual abuse as a factor in adolescent pregnancy and child maltreat- ment. Fam Plan Perspect. 1992:24:4-1 i.

35. Briere J. Melhodoiogical issues in the study of sexuai ahuse effects. J Consult Clin Psychot. 1992:6(1:196-203.

36. Cicchetti D. Rizley R. Deveiopmentai perspectives on the etiolog\'. inlergenera- tional transmission, and sequelae of child maltreatment. New Dir Child De\'. 1981;!!: 31-55.

37. Yarrow MR. Campbell JD. Burton RV. Recollections of chiidhixjd: a sludy of the retrospective method. Monogr Soc Res Child Dev \'il{)•.\^S(\:^H, 5).

.38. Widom CS. Sampling hiases and implica- tions for ehild abuse research. Am J Orthopsychiatiy. 1988:58:260-271).

39. Newberger EH. Reed RB. Daniel JH. Hyde JN. Kotelchuck M. Pediatric social iliness: toward an etioiogical ciassiHcation. Pediatrics. 1977;5(): 178-185.

40. Straus MA, Gelle.s RJ, eds. f^ysical Vio- lence in American Families: Risk Ftictors and Adaptations to Violence in iHN.'y Familie.s. New Brunswick. NJ: Transaction Pulish- ers, 1990.

41. Widom CS. Child ahuse. neglect and adult behavior: research design and tindings on criminality, violence, and child abuse. Am J Onhopsvcliiatry. 1989:59:355-367.

42. Beitchman JH. Zucker KJ, Hood JE. daCosta GA. Akman D. A review of the short term effects of chiidhixxJ sexual abuse. Cltild Abu.se Negl. 199i;l5:537-556.

43. Leventhal JM. Research strategies and methodoiogic standards in studies of risk

factors for child abuse. Child Abuse Negt. 1982:6:113-123.

44. Leibowitz A, Eisen M. Chow WK. An economic model of teenage pregnancy decision-making. Demographv. 19S6;23:67- 77.

45. Cooksey EC. Factors in the resolution of adolescent premarital pregnancies. Demog- raphy. 199(l;27:2()7-218.

46. Cixiksey EC. Outcomes of Adolescent First Premarital Pregnancies: The Injiuence of Family Background. Providence. RI: Brown University: 1988. DissertaUon.

47. Fisher B. Weisberg DK, MaroUa T. Report on Adolescent Mole Prostitution. San Fran- cisco. Calif: Urhan Rural System Associ- ates: 1982.

48. Widom CS. ChildhtMid victimization and risk for adoiescent problem hehaviors. In: Uimh ME. Ket(er!inus R. eds. Adolescent Problem Behaviors. New York. NY: Eribiium. 1994.

49. Witlom CS. The cycle of vioience. Science. 1989;244:i6()-166.

5(1. Inciardi JA. Trading sex for crack among juvenile drug users: a research note. Con- temp Dmg Probl Winter 1989:684-71 Kl.

51. Inciardi JA, Pottieger AE. Forney MA, Chitw<x»d DD. McBride DC. Prostitution. IV drug use. and sex-for-crack exchanges among serious delinquents: risks for HIV infection. Criminology. 199i:29:221-235.

52. Kuhns JB. Heide KM. Siiverman I. Sub- stance use/misuse among iemale prosti- tutes and femaie arrestees. Int J Addict. 1992;27:128.1-1292.

53. Philpot CR, Harcourt CL, Edwards JM. Drug use by prostitutes in Sydney. Br J Addict. 1989:84:499-505.

54. Marshall N. Hendllass J. Drugs and prosti- tution..//M/^/.ww.v. Spring 1986:227-248.

55. Goldstein PJ. Prostitution and Drugs. Toronto. Ontario, Canada: tX' Heath & Co: 1979.

56. Cottier LB. Helzer JE. Tipp JE. Lifetime patterns of substance use among generai popuiation subjects engaging in high risk sexuai hehaviors: implications for HIV risk. AmJDmgAlcoholAhu.se. 1990:16:207-222.

1612 American Journalof Public Health November 1996, Vol. 86, No. 11