Sexual abuse includes a spectrum of activities ranging from rape to physically less intrusive sexual abuse.^1,2 Child sexual abuse has been a topic of public and governmental investigation for the past few years in Korea because of the notorious child sexual abuse victim case called the 'Cho Doo-Sun case' named after the assaulter in 2008. Since then Koreans have become more concerned about child sexual abuse, the issue has come under public scrutiny, and the Korean National Statistical Office reported more than 1,017 child sexual abuse cases among children under 13 years old in 2009. Sexually abused children shows sexual behavior problems and other generic behavior problems^3,4 and are prone to psychiatric disorders such as depression, post-traumatic stress disorder, dissociative disorder, and so forth as a result.⁵ Sexually abused children as a group exhibit more sexual behavior problems than non-sexually abused children.^6-8 Because of incomplete cognitive and language development, young children are not able to disclose sexual abuse, and caretakers must have knowledge of the symptoms of child sexual abuse to recognize the abuse as early as possible and provide them with more appropriate help. Therefore mental health professionals have developed and used standardized assessment measures as a need of accurate evaluating the symptoms of child sexual abuse. There are several instruments designed specifically for children and adolescents who are suspected of having been or being currently sexually abused, which are the Anatomical Doll Questionnaire (ADQ), the Checklist of Sexual Abuse and Related Stressors (C-SARS), and the Child Sexual Behavior inventory (CSBI), Negative Appraisals of Sexual Abuse Scales (NASAS), and Sexual Abuse Fear Evaluation (SAFE).^9-11 The ADQ is semi-structured 2-7 years old child interview by the use of anatomical dolls (other than verbal format). It is purposed to find out type of abuse, to observe child's behaviors and perceptions of interview quality. The C-SARS is self-report, assessing both reports and degree of stressful events associated with a sexual abuse. The CSBI assesses a wide range of sexual behavior that covers a number of domains including boundary problems, exhibitionism, gender role behavior, self-stimulation, sexual anxiety, sexual interest, sexual intrusiveness, sexual knowledge, and voyeuristic behavior by caregiver. The NASAS measures perception of threat of harm associated with sexual abuse. The SAFE assesses abuse-related fear. There are other measures evaluating children's behavior by caregiver, which are Child Behavior Checklist (CBCL), Trauma Symptom Checklist for Young Children (TSCYC). However these are not for specific subject who are sexually abused children. The CSBI was developed following research that identified the sexual items on the Child Behavior Checklist (CBCL) developed by Achenbach et al.¹² The CSBI was developed by Friedrich in 1991. Two other versions have been developed since the first version of the CSBI.¹³ The CSBI has demonstrated adequate reliability and validity.^13,14 Koreans has been directly influenced by Confucianism since the Chosun dynasty (14th-19th centuries) and the word 'sexual' has been a taboo, and which has impeded the inquiries in children's experiences of sexual abuse directly and even indirectly. Even if external evidence of abuse such as gonorrhea had existed, disclosure rates among children have been less than 50%.¹⁵ As a result, recent research on the consequences of sexual abuse has focused on the need for a more systematic examination,⁹ especially by caretakers. The CSBI has advantages of sexual abuse-specific measure and being useful for evaluating sexual behavior of children who have been sexually abused or suspected of having been sexually abused.¹⁶ Until now, little research for measuring child sexual behavior after sexual abuse has been carried out in Korean society.

The aim of this study is to evaluate the reliability and validity of the Korean version of the Child Sexual Behavior Inventory (CSBI) in Korean children aged from 6 to 12 years old and the suitability of and potential for clinical application of the CSBI in a Korean population.

This study examined a range of sexual behaviors measured by the Korean version of the Child Sexual Behavior Inventory (CSBI) in a sample of Korean children aged 6 to 12 years old. The CSBI is specific measure for sexually abused children and has advantages of evaluating children who are incomplete in terms of cognition and language development because it is reported by caregiver. The purpose of this study was to examine the psychometric properties including the reliability and validity of the tool, and to evaluate the possibility of clinical application of the Korean version of the CSBI in Korean society.

There was a higher non-response rate on the demographic data in the sexually abused children (SA) group (maternal educational level 29.5%, paternal educational level 30.4%, economic status 29.5%) than the typically growing children (TG) group (maternal educational level 7.0%, paternal educational level 0.6%, economic status 1.3%). The economic status in the SA group was lower than was that of the TG group. The parental education level was also lower in the SA group than in the TG group.

The author of the English version of the CSBI reported that the alpha coefficient of that version of CSBI was 0.72 for the normative sample and 0.92 for the sexual abuse sample.¹⁴ The alpha coefficient was higher in the SA group than the TG group in both our study and the original study.¹⁴ This indicates that the Korean version of the CSBI, like the original scale, is more reliable measurement for children with sexual abuse.

In the TG group, the alpha coefficient was low as 0.47 because 16 items all with a score of '0' were excluded from the calculations. In explaining the large number of '0' answers, we must take into consideration the conservative Korean culture based on Confucian strictures including sexual taboos even among the adults, which likely affected the parents' ability to observe and report their children's sexual behavior. Because the CSBI was designed to detect problematic sexual behavior earlier among sexually abused children, the Cronbach's alpha value for the SA group of 0.83 and that of 0.84 for all of the participants taken together showed the good reliability of this scale. If item 38 was deleted, the Cronbach's alpha value increased to 0.85. Item 38 is an essay question, and a number of participants did not fill out this item. This was likely to have contributed to the elevation of Cronbach's alpha when item 38 was deleted. The degree of the increase is so slight that item 38 did not need to be excluded.

Sexually abused children had higher scores than nonabused children in the sex problems subscale of the CBCL in previous studies; however, the sex problems subscale scores of the CBCL were not significantly different between the two groups.^22,26,27 The sex problems subscale of the CBCL consists of only 6 items²⁸ and it does not discriminate the problematic sexual behavior from other problematic behavior in the sexually abused children, which implies that the CSBI, more reliable and valid instrument, is strongly necessary in Korean culture. In our study, the total score of the CSBI was more strongly correlated with the total problematic behavior (r=0.43) and internalizing problems (r=0.40) subscales rather than the sex problems (r=0.27) and externalizing problem (r=0.38) subscales. However, in the study of Friedrich et al., the CSBI total score showed stronger correlation with the CBCL's externalizing problems subscale (normative group r=0.43, sexual abuse group r=0.47) than the internalizing problems subscale (normative group r=0.35, sexual abuse group r=0.36).

Significant group differences of in gender, parental educational level, and economic status were identified both in our study and the study of the original author.¹⁴ Our finding that the CSBI total score was significantly higher in the SA group than the TG group not only for the total sample but also for the two age groups (6 to 9 and 10 to 12-years old) considered separately according to developmental stage accords with the study of the original author.

In this study, the mean total score of the CSBI was lower than that of the original author's study, not only in TG group (0.43±1.31 vs. 3.7±4.5), but also in SA group (4.36±6.47 vs. 13.5±15.2).¹⁴ These results reflect the characteristic of Korean society that people generally want to provide socially desirable answers, even more than in Western society. More '0' scores on more items also lowered the total score. These results corresponded with previous studies showing that in Korean society people generally want to provide socially desirable answers, even more than in Western society.³⁰

Comparisons of all 38 items of the CSBI between the two groups were also examined. Friedrich et al.¹⁴ found that all 38 items of the CSBI were endorsed significantly more often (p<0.05) for the sexually abused children samples than for the normative samples. The present study revealed that some items (item 6, 8, 9, 10, 18, 22, 28, 29, 30, 31, 33, 34, and 38) of the CSBI did not differ significantly across the two groups. It cannot be completely ruled out that Korean children show fewer sexual behaviors after sexual abuse than Western children. Another explanation might be a cultural difference between Eastern and Western countries. Cho et al.²⁹ suggested that the stigma for mentally ill patients was so severe in Korea that many of the respondents had a tendency to conceal their symptoms.³⁰ This interpretation is supported by the finding that the means of the CSBI total score of not only SA group but also TG group in our study are lower than those of previous studies with Western samples. Taylor et al.³¹ explained that Korea is one of the countries in which subjects show a response bias because they want to give an answer as socially desirable as possible. Thus the nuance or meaning of words used in the CSBI scale concerning sexual behavior could repulse Korean participants. However, most sexual behaviors indicated by these items have been regarded as important in child and adolescent psychiatry.^32,33 Therefore, although there were no significant group differences for these individual items, they still have clinical value, and we must apply the results of all of the items to effectively evaluate and treat sexually abused children.

Item 14 and item 20 were scored as '0' by all 280 participants. Item 14 is "Touches animals' sex parts" and item 20 is "Pretends that dolls or stuffed animals are having sex." Most Koreans live in apartments, unlike many Westerns, and apartment dwellers generally prohibited from keeping animals. Thirtynine percent of U.S. households own at least one dog and thirty-three percent of U.S. households own at least on cat, on the other hand, approximately twenty-one percent of Korean households own pets.^34,35 Therefore, Korean children tend to have less opportunity to touch animals. However, Item 20 is commonly reported sign of child sexual abuse in Western society, so we need to evaluate more information about this item through large scale research henceforth.

There are limitations to our study. First, in this study, testretest reliability was not examined and a later study should investigate the test-retest reliability of the Korean version of the CSBI to confirm the reliability of this scale with greater accuracy. Second, sample size of this study was quite small in comparison with the study of previous authors.¹⁵ We recruited 158 participants for the TG sample and 122 for the SA sample. Third, although the CSBI was developed for measuring sexual behavior in children aged 2 to 12 years, we only evaluated children ranging from 6 to 12 years the analysis of children aged 2 to 5 years old was absent from our study. Fourth, there was selection bias in including only sexually abused children identified to be in need of mental health services. Fifth, the study was not intended to measure the underlying psychiatric disorders of the SA group participants. Previous or comorbid psychiatric disorders may influence sexual behaviors. Sixth, the CSBI scoring system contains Developmentally Related Sexual Behavior (DRSB) and Sexual Abuse Specific Items (SASI) according to gender and age in addition to the CSBI total. We did not examine these two scales. The DRSB and SASI could discriminate among children's various symptoms to distinguish normal developmental sexual behavior and sexual abuse-related behavior. Thus, it is helpful that psychiatrists or practitioners understand a child's age-appropriate sexual behavior.

In conclusion, the Korean version of the Child Sexual Abuse Inventory (CSBI) is reliable and valid, and it can be an appropriate clinical tool for assessing the sexual behavior of Korean children aged from 6 to 12 who are suspected to have been sexually abused.