Child Sexual Behavior Inventory
- Parallel/Alternate Forms
- Translation Quality
- Population Information
- Pros & Cons/References
Friedrich, W.N. (1997). Child Sexual Behavior Inventory: Professional Manual. Odessa, FL: Psychological Assessment Resources.
The 38-item Child Sexual Behavior Inventory (CSBI) was developed to assess children who have been sexually abused or are suspected of having been sexually abused. The measure is designed to be completed by a female caregiver. It is one of the most widely used measures of sexual behaviors.
It yields a total CSBI score, a Developmentally Related Sexual Behavior Score, and a Sexual Abuse Specific Items Score, with
norms by age and gender for these scales.
It also yields scores on 9 domains: 1) Boundary Problems, 2) Exhibitionism, 3) Gender Role Behavior, 4) Self- Stimulation, 5) Sexual Anxiety, 6) Sexual Interest, 7) Sexual Intrusiveness, 8) Sexual Knowledge, and 9) Voyeuristic Behavior.
The CSBI is a revision of the CSBI-R and CSBI-1. It contains 22
of the items from the previous version, with the remaining items reworded for greater readability (Friedrich et al., 2001).
The response format for each behavior is a 4-point Likert scale, which indicates the frequency of behavior (0=never to 3=at least once per week).
|Trauma: Sexual abuse||Sexual Behavior||not available|
|Verbal Sexual Behavior||not available|
|Sexual Intrusiveness||not available|
|Gender Role Behavior||not available|
|Boundary Issues||not available|
|Sexual Interest||not available|
1. There are two previous versions of this measure (versions 1 and 2). The CSBI-2 included 9 items that were revised or added to the original version.
The 36-item CSBI-2 can be seen at the U.S. Department of Health and Human Services, National Clearinghouse on Child Abuse and Neglect Information
The CSBI differs from the CSBI-2 in that it contains 22 of the original items, and the remaining items were reworded to make them easier to read. The manual reports that the latest version also differs from previous versions in terms of the age groups for which norms were calculated.
Earlier versions had norms for ages 2-6 and 7-12. The latest version has norms for 2-5, 6-9, and 10-12.
2. There is an Adolescent Clinical Sexual Behavior Inventory (ACSBI), a 45-item Self- and Parent-Report measure specific to the age-appropriate sexual behaviors of adolescents. (Friedrich, Lysne, Sim, & Shamos, 2004).
From the above study, it appears that this measure assesses children between the ages of 12-18 years. Self-Report and Parent-Report versions are reviewed in this database.
1. Normative data for the CSBI represent 1,114 children, combined from three nonclinical samples: (a) 723 children who were in the waiting area of a
Community Pediatrics Clinic in Rochester, Minnesota; (b) 111 children who were in the waiting area of a Community Family Medicine Clinic in
Rochester, Minnesota; and (c) 280 children from the Los Angeles, California, area, the majority of whom were lower income and of minority status."
The total sample included the following: 49.7% female, 51.3% males; 77.7% Caucasian, 7.7% African-American, and 11.6 % Latino.
Norms are presented in the manual’s appendix by gender and age group (2- 5 years, 6-9 years, 10-12 years).
There are no norms for fathers.
2. Data were also collected in multiple sites in the U.S. and Canada on 512 children with a documented history of sexual abuse. The average time since abuse was 10.9 months (SD=13.9), the average length of abuse was 9.3 months (SD=12.3 months). Children were aged 2-12 (M=7.44; SD=2.63); 62.7% female, 37.3% male; 76% White, 6.9% African American, 8.7% Hispanic, 2.2% Asian, 2.4% Native American, 3.8% Other; 62.4% of families had annual incomes below $25,000.
|Test-Retest- # days: 14||Acceptable||r||0.91|
Friedrich et al. (2001) present reliability data on three samples: 1) 1,114 children who comprised the normative sample, 2) 620 sexually abused children, and 3) 577 psychiatric outpatients. Similar data are presented in the manual, although no outpatient data are presented. All data reported below and in the table are from the current version of the CSBI.
63 parents from the normative sample completed a second CSBI two weeks later.
Examined in multiple samples (Friedrich et al., 2001). The alpha for the sexual abuse sample is reported above.
Normative sample: whole sample (alpha=.72); males (.72), females (.73) Children aged 2-5 (.72), aged 6-9 (.71), aged 10-12 (.71)
Sexual abuse sample: whole sample (alpha=.92); males (.92), females (.93) Children aged 2-5 (.93), aged 6-9 (.91), aged 10-12 (.91)
Outpatient sample: whole sample (alpha=.83); males (.79), females (.88)
There were no age differences when internal consistency was examined separately for ages 2-5, 6-9, and 10-12 years.
For correlations between 24 normative sample mother-father pairs, married and living together (r=.79, p<.01).
For correlations between 22 primary nurses on an inpatient psychiatric ward (slightly modified CSBI) and primary female caregivers (r=.42, p<.01).
For correlations between 61 teachers (using the 3 sex-related items of the TRF) and parents (r=.44, p<.01).
From Friedrich (1997, p. 37)
A pool of 40 items was developed after it was determined that sexual behavior items on the CBCL were useful in discriminating sexually abused from nonabused children.
Items were piloted in a pilot study involving 71 nonabused and 35 sexually abused children, with mean scores differing between the two groups.
Interviews with 32 caregivers of sexually abused children led to the identification of additional sexual behaviors; 8 items were added and 16 were revised.
Following additional data collection, 13 of the 48 items were dropped (5 were poorly worded,3 pertained to the child’s exposure to family nudity, and 5 assessed toileting or somatic behaviors), leaving 35 items.
The 35-item CSBI was studied and found to differentiate sexually abused from nonabused children. It was later revised (6 items were dropped, 3 were reworded, and 7 new items were added).
The 36-item CSBI was found to also differentiate sexually abused from nonabused children. It was revised to become the current version. It consists of 22 of the original items and 12 of the items “rewritten for clarity and simplicity.” One item was dropped and 3 were added.
|Validity Type||Not known||Not found||Nonclincal Samples||Clinical Samples||Diverse Samples|
|Sensitive to Change||Yes|
|Sensitive to Theoretically Distinct Groups||Yes||Yes|
Numerous studies using different versions of the CSBI have demonstrated the validity of the different versions. CSBI scores are positively correlated with scores on the Child Behavior Checklist (Drach, Wientzen, & Ricci, 2001; Mannarino & Cohen, 1996a), intensity of parents’ reaction to the abuse
(Mannarino & Cohen, 1996b), and teacher ratings of sexual behavior on the Teacher’s Report Form (Friedrich et al., 1992).
The CSBI discriminates between sexually abused and non-sexually abused children (Friedrich, Jaworski, Huxsahl, & Bengtson, 1997; Wherry, Jolly,
Feldman, & Adam, 1995).
It has also been found to be sensitive to treatment effects with sexually abused children aged 3-6 treated using Cognitive Behavioral Therapy, showing greater declines on this measure than with children treated using nondirective supportive therapy (Cohen & Mannarino, 1996). More details regarding the validity of earlier versions of the CSBI can be found in the manual (Friedrich, 1997).
For the current version of the CSBI, Friedrich et al. (2001) examined the psychometrics of the measure using normative, outpatient, and sexual-abuse samples. CSBI scores were significantly correlated with CBCL Internalizing and Externalizing scores for each of the three groups studied. They found that the sexual-abuse group scored higher on all items than did the other two groups, and the outpatient sample scored significantly higher than did the normative sample on 7 items. Analyses by age and gender indicated that this pattern was replicated by age and gender.
CSBI scores were also significantly related to aspects of the abuse, such as penetration, duration, frequency, multiple perpetrators. Drach et al. (2001), however, found that in their sample of children involved in forensic evaluations, there was no relation between CSBI scores and what they determined a “diagnosis” of sexual abuse (see Notes under “Criterion Validity”). However, this may have been due to how they made their “diagnosis.”
STUDIES WITH OTHER CULTURAL GROUPS AND DIVERSE POPULATIONS
Note: These studies have generally been conducted with earlier versions of the
Studies have found that children in other cultural groups exhibit different levels of sexual behaviors. In a samples of non-sexually abused children, Larsson, Svedin, & Friedrich (2000) and Friedrich et al. (2000) found that Swedish and Dutch children exhibit more sexual behavior than did American children.
The authors suggested that cultural context affects what behavior is permitted and what behavior is considered problematic. These studies suggest that different norms would be needed for different cultural groups.
Schoentijes, Deboutte, & Friedrich (1999) examined the psychometrics of the 44-item CSBI in a normative Dutch sample. They report the frequency with which items were endorsed. They found good internal consistency (alpha=.86), and
identified 7 factors using principal components analysis with a varimax rotation.
CSBI scores were related to CBCL Internalizing and Externalizing scores.
|Not Known||Not Found||Nonclinical Samples||Clinical Samples||Diverse Samples|
Friedrich et al. (2001) report the Sensitivity and Specificity rates shown above. The rates above are the average rates of the normative versus sexually abused group.
.88-.95 for comparisons of normative versus sexually abused groups
.43-.55 for comparisons of psychiatric outpatient versus sexually abused groups
.43-.55 for comparisons of normative versus sexually abused groups
.78-.87 for comparisons of psychiatric outpatient versus sexually abused groups.
Drach, Wientzen, & Ricci (2001) report a Positive Predictive Value of .28 and Negative Predictive Value of .72 in a forensic child sexual-abuse population.
These data are not shown in the table because there are methodological problems in the way that they defined a “diagnosis of sexual abuse.” They defined the “diagnosis” as the result of the forensic evaluation.
They suggested that while positive screens should be followed up they should not be used as a diagnostic indicator. The authors did suggest that the way they determined “diagnosis” may yield too many false negatives.
1. PRO: The authors conducted age and gender analyses of the current CSBI, looking at differences between boys and girls, aged 2-5, 6-9, and 10-12. The results generally show similar reliability and validity data across all groups.
2 While the psychometrics have been examined in studies involving lower socioeconomic status individuals, the majority of studies have been conducted with Whites.
3. The measure was designed to be completed by female caregivers. Although the testretest reliability between mothers and fathers from intact homes is good, more research is needed regarding reports from fathers if fathers’ reports are to be used.
4. While the measure has been translated in multiple languages (see Languages Other Than English), translations and studies that involved them were generally conducted with earlier versions of the CSBI.
|Language:||Translated||Back Translated||Reliable||Good Psychometrics||Similar Factor Structure||Norms Available||Measure Developed for this Group|
The clinical sample included 276 children aged 2-12 with confirmed history of sexual abuse, typically occurring within the last 12 months. Sample was multisite (East and West coasts, Midwest, major communities). Most children were referrals to agencies; others were in therapy at the time of the study.
Data regarding the normative and clinical samples used to validate and standardize the current version of the CSBI are provided in the Notes under “Norms.”
|Population Type:||Measure Used with Members of this Group||Members of this Group Studied in Peer-Reviewed Journals||Reliable||Good Psychometrics||Norms Available||Measure Developed for this Group|
|1. Developmental disability|
|3. Lower socio-economic status||Yes||Yes||Yes||Yes|
|4. Rural populations|
Pros & Cons/References
1. Sexual behaviors are important to assess, and this measure appears useful in capturing different dimensions of sexual behaviors.
2. This is one of the most widely used and well-researched tests tapping the domain of child sexual behaviors. The Trauma Symptom Checklist for Children also assesses for sexual behavior, but does so using self-report.
3. Wording of items: items appear to be clear. Rating scale is an improvement over previous rating scales with sexual behavior items (e.g., CBCL).
4. Psychometrics have been evaluated by age and gender.
5. Measure has been demonstrated to show change with treatment.
1. The item uses a 6-month timeframe. This timeframe is not appropriate for treatment outcome studies of shorter or similar duration. The timeframe would add error if it is adhered to and informants are responsive to it. Although the timeframe can be adjusted, the norms were gathered using that time period and would not necessarily apply with a new timeframe.
2. The measure is face valid and there are no validity scales. There were validity items used during the normative studies, but they were dropped from the current version. The manual advises users to determine, during follow-up interviews, whether caregivers read and interpreted items appropriately.
3. Although psychometrics have been examined in studies involving lower-socioeconomicstatus individuals, the majority of studies have been conducted with Whites. More studies are needed with diverse populations.
4. The measure was designed to be completed by female caregivers. Although the testretest reliability between mothers and fathers from intact homes is good, more research is needed regarding reports from fathers if fathers’ reports are to be used.
5. While the measure has been translated in multiple languages, translations and studies
that involved them were generally conducted with earlier versions of the CSBI.
6. THIS IS A CAUTION RATHER THAN A CON: While the measure has been found to be valid and useful, the authors and others point out that many sexually abused children display low levels of sexual behavior problems, and many children who were not sexually abused exhibit high levels of sexual behavior problems.
The reference for the manual is:
Friedrich, W.N. (1997). Child Sexual Behavior Inventory: Professional Manual. Odessa, FL: Psychological Assessment Resources.
A PsychInfo literature search of "Child Sexual Behavior Inventory” or “CSBI" anywhere (6/05) revealed that the measure has been referenced in 88 peer-reviewed journal articles.
Note: It is difficult to search only for this version, but a search for the manual for this version revealed it had been referenced in 11 peer-reviewed journal articles. Given that this most likely underrepresents the measure’s use, the numbers from the full search are presented.
A sampling of the articles is listed below along with references for the original versions of the CSBI:
Cohen, J.A, & Mannarino, A.P. (1996a). Factors that mediate treatment outcome of sexually abused preschool children. American Academy of Child & Adolescent Psychiatry, 35(10), 1402-1420.
Cohen, J.A., & Mannarino, A.P. (1996b). A treatment outcome study for sexually abused preschool children: Initial findings. Journal of the American Academy of Child and Adolescent Psychiatry, 35(1), 42-50.
Costintino, C.E., Meyer-Bahlburg, H.F.L., Alpert, J.L., Weinberg, S.L., & Gaines, R.G. (1995). Sexual behavior problems and psychopathology symptoms in sexually abused girls. American Academy of Child & Adolescent Psychiatry, 34(8), 1033-1042.
Drach, K.M., Wientzen, J., & Ricci, L.R. (2001). The diagnostic utility of sexual behavior problems in diagnosing sexual abuse in a forensic child abuse evaluation clinic. Child Abuse & Neglect, 25, 489-503.
Friedrich, W.N. (1993). Sexual behavior in sexually abused children. Violence Update, 3(5), 7-11.
Friedrich, W.N., Grambsch, P., Broughton, D., Kuiper, J., & Beilke, R.L. (1991). Normative sexual behavior in children. Pediatrics, 88(3), 456-464. (Reference for CSBI-1.)
Friedrich, W.N., Grambsch, P., Damon, L., Hewitt, S.K., Koverola, C., Lang, R.A., Wolfe, V., & Broughton, D. (1992). Child Sexual Behavior Inventory: Normative and Clinical Comparisons, Psychological Assessment, 4(3), 300-311.
Friedrich, W.N., Jaworski, T.M., Huxsahl, J.E., & Bengtson, B.S. (1997). Dissociative and sexual behaviors in children and adolescents with sexual abuse and psychiatric histories, Journal of Interpersonal Violence, 12(2), 155-171.
Friedrich, W.N., Lysne, M., Sim, L., & Shamos, S. (2004). Assessing sexual behavior in high-risk adolescents with the Adolescent Clinical Sexual Behavior Inventory. Child Maltreatment, 9(3), 239-250.
Friedrich, W.N., Sandfort, T.G.M., Oostveen, J., & Cohen-Kettenis, P.T. (2000). Cultural differences in sexual behavior: 2-6 year old Dutch and American children. Journal of Psychology and Human Sexuality, 12, 117-129.
Larsson, I., Svedin, C., Friedrich, W.N. (2000). Differences and similarities in sexual behavior among pre-schoolers in Sweden and USA. Nordic Journal of Psychiatry 54,4:251- 258.
Mannarino, A.P., & Cohen, J.A. (1996a). Abuse-related attributions and perceptions, general attributions, and locus of control in sexually abused girls. Journal of Interpersonal Violence, 11(2), 162-180.
Mannarino, A.P., & Cohen, J.A. (1996b). Family-related variables and psychological symptom formation in sexually abused girls. Journal of Child Sexual Abuse, 5(1), 105-120.
National Clearinghouse on Child Abuse and Neglect, U.S. Department of Health & Human Services, 2001.
Schoentjes, E., Deboutte, D., & Friedrich, W. (1999). Child Sexual Behavior Inventory: A Dutch-speaking normative sample. Pediatrics, 104(4), 885-893.
Silovsky, J.F., & Larissa, N. (2002). Characteristics of young children with sexual behavior problems: A pilot study. Child Maltreatment, 7(3), 187-197.
Wherry, J.N., Jolly, J.B., Feldman, J., & Adam, B. (1995). Child Sexual Behavior Inventory scores for inpatient boys: An exploratory study, 4(3), 95-105.