**4. Childhood abuse assessment questionnaires**

To encourage the practitioners to further explore the possibility that an abuse history may exist in a given patient, a number of time efficient childhood abuse instruments have been developed to facilitate this objective. Unfortunately, the vast majority of the childhood trauma questionnaires reviewed provide either no or limited psychometric support. Of the numerous questionnaires reviewed, seven were selected for review which have been frequently adopted and have provided adequate psychometric support. They are as follows: (1) The Rape Aftermath Symptom Test; (2) The Scarlett O'Hara VMMPI Configuration; (3) The Child Abuse and Trauma Scale; (4) The Childhood Trauma Questionnaire; (5) The Trauma Symptom Checklist; (6) The Trauma Symptom Inventory; and (7) The Binghamton Childhood Abuse Screen.

The first two of the above inventories were not specifically designed to assess childhood abuse, but rather were designed to assess events frequently associated with a history of abuse (e.g., rape and parental alcoholism). The remaining five questionnaires were specifically designed to assess an abusive childhood history. In the conduction of this review, I would like to acknowledge the review contribution provided by two of my former students, Timothy Lock (1997) and Brian Castelda (2003).

The format adapted for these reviews is to first provide a basic description of the questionnaire, followed by an assessment of the inventories reliability, criterion related validity, convergent validity, discriminate validity, and criterion validity.

1. The Rape Aftermath Symptom Test (RAST)

This test constructed by Kirpatrick (1988) is a 70 item self report inventory comprising items designed to assess the presence of psychological symptoms and potentially fear producing stimuli. The RAST was developed using a limited sample size of 137 rape survivors and 139 non-abused victims. According to standards suggested by Devellis (1991), a sample size greater than 300 participants is needed. Nevertheless, the RAST has provided good testretest reliability with non-rape victims (r = .95 at 2.5 month intervals) and good internal consistency reliability (α = .95). Criterion related validity was determined by discriminating statistically between scores of non-rape victims and rape victims at 6 to 21 days, 3 months, 6 months, and at 1, 2, and 3 years past the rape experience (Kirpatrick 1988; Resnick, Kirpatrick, & Lizovashky, 1991).

The psychometric properties of the RAST could be improved by increasing its validation size and by establishing convergent and discriminatory validity. Since the RAST, items are related to a rape experience, it directs relevance to the assessment of childhood abuse is somewhat limited.

#### 2. The Scarlett O'Hara VMMPI Configuration

6 Sexual Abuse – Breaking the Silence

this author's opinion that if abuse trauma memories are forthcoming, one should never agree they are factual. This responsibility must be left up to patients to decide what's true and what's not true. The key responsibility of the therapist is to determine whether the release and extinction of the affect associated with a reported trauma leads to therapeutic gains. It has been my extensive experience (30 years) from providing treatment for adult survivors of childhood abuse that it does, a finding I am sure can be confirmed by many other therapists who treat abused patients. Finally, Briere's (1992b) conclusion is correct that most researchers in the area of child abuse agree that only a small proportion of people who

Although the establishment of external corroboration of reported abuse memory recovery is desirable, it is difficult to achieve (Briere, 1992a), some success in achieving this objective has been reported (e.g., Herman & Schatzow 1987); Feldman-Summers & Pope 1994). For example, Herman & Schatzow (1987) had a group therapy treatment of CSA survivors and encouraged them to obtain corroborating information from internal sources (e.g., sibling, family memories, medical records, court proceedings, etc.). They found that 39 out of their 53 group participants (74%) found some external confirmatory evidence. Although such evidence is not definitive, it represents a first step in the attempt to provide some form of

To encourage the practitioners to further explore the possibility that an abuse history may exist in a given patient, a number of time efficient childhood abuse instruments have been developed to facilitate this objective. Unfortunately, the vast majority of the childhood trauma questionnaires reviewed provide either no or limited psychometric support. Of the numerous questionnaires reviewed, seven were selected for review which have been frequently adopted and have provided adequate psychometric support. They are as follows: (1) The Rape Aftermath Symptom Test; (2) The Scarlett O'Hara VMMPI Configuration; (3) The Child Abuse and Trauma Scale; (4) The Childhood Trauma Questionnaire; (5) The Trauma Symptom Checklist; (6) The Trauma Symptom Inventory; and (7) The Binghamton

The first two of the above inventories were not specifically designed to assess childhood abuse, but rather were designed to assess events frequently associated with a history of abuse (e.g., rape and parental alcoholism). The remaining five questionnaires were specifically designed to assess an abusive childhood history. In the conduction of this review, I would like to acknowledge the review contribution provided by two of my former

The format adapted for these reviews is to first provide a basic description of the questionnaire, followed by an assessment of the inventories reliability, criterion related

This test constructed by Kirpatrick (1988) is a 70 item self report inventory comprising items designed to assess the presence of psychological symptoms and potentially fear producing stimuli. The RAST was developed using a limited sample size of 137 rape survivors and 139 non-abused victims. According to standards suggested by Devellis (1991), a sample size

describe abuse experiences make them up.

**4. Childhood abuse assessment questionnaires** 

students, Timothy Lock (1997) and Brian Castelda (2003).

1. The Rape Aftermath Symptom Test (RAST)

validity, convergent validity, discriminate validity, and criterion validity.

external corroboration.

Childhood Abuse Screen.

Goldwater & Duffy (1990) found that their Minnesota Multiphasic Inventory (MMPI) V configuration is suggestive of parental alcoholism which is frequently found in the history of childhood abused victims. Their "V" configuration refers to the MMPI scales 4 and 6 being elevated with T-scores of 65 or above, and scale 5 falling at least 30 T-scores below scales 4 and 6.

The test-retest reliability of the subscales used in the Scarlett O'Hara V configuration are as follows: Scale 4 r = .79, α = .60, 5 r = .73; α = .37; scale 6 r = .58, α = .34 (Hathaway McKinley, 1989). Criterion related validity was determined by the test ability to discriminate between participants with a history of abuse (or an alcoholic parent) and non-abuse. Histories of abuse or parental alcoholism were extracted from patients' hospital charts after discharge. The sample size evaluated consisted of 79 adult female inpatients. A key advantage of the Scarlett O'Hara V MMPI configuration is the ability to add additional information on personality structure and psychopathology. However, the test was developed using a small and restricted sample size. Further, the test-retest and/or internal consistency reliability for some of the scales are lower than desired. Finally, the presence of a parental history of alcoholism does not in and of itself provide any direct confirmation that the parental offspring have been subjected to an abusive childhood history. It simply represents a hypothesis in need of further confirmation.

3. The Childhood Abuse and Trauma Scale (CAT)

This questionnaire, developed by Sanders & Becker-Lausen (1995), was initially created as a research scale but later it was considered to be a useful tool in clinical assessment as an initial screening instrument. The scale consists of thirty-eight items to assess various forms of childhood physical, sexual, and maltreatment abuse. Part of the research objective of the scale was to combat respondent tendencies toward giving socially desirable responses. To achieve this objective, the scale was made up of general questions about the frequency of different childhood abuse experiences while allowing the respondent to determine his own evaluation of the severity of the stress experiences. Internal consistency of the entire CAT, as reflected in Cronbade's alpha was established to be .90. Test-retest reliability, determined over 6-8 weeks intervals, was .89. Besides measures of reliability, convergent validity of the CAT has been demonstrated in establishing correlations with dissociation (r = .29). The CAT also scored higher when a group diagnosed with Multiple Personality Disorder was compared to a normative sample (Sanders & Becker-Lausen, 1995).

A Review of Childhood Abuse Questionnaires and Suggested Treatment Approaches 9

the correlation between two forms of almost identical abuse measures and points out that these correlations tell us little of the questionnaire's ability to discriminate abuse victims. Bursten also criticized Bernstein et al. (1994) study for the author's use of the CIQ in their validation study because it was experimental and undergoing validation. Lastly, Bursten notes that aside from adolescent psychiatric patients, no normative data exists for the CTQ that would allow for its use in a clinical setting. Nevertheless, the CTQ shows considerable

The TSC-40 appears to be the most popular childhood abuse questionnaire available (Castelda, 2003) perhaps in part because of extensive research history and strong psychometric support. TSC-40 was developed from the Trauma symptom Checklist – TSC-33 (Briere & Runtz, 1989) which was developed to assess the impact of trauma especially the long-term effects of CSA. The emphasis was on gaining a more complete understanding of the precise patterns of symptomatology associated with a history of CSA (Gold, Milan, Mayall & Johnson, 1994). Participants respond to items by rating how often they had experienced each in the last two months (0 = never, 4 = very often. In a sample of woman at a crisis intervention clinic, TSC-33 was found to be internally consistent, alpha = .89 capable of correctly identifying 79% of self-reported CSA survivors. Twenty-three of these items discriminated between abused and non-abused participants. The TSC-33 consists of five subscales labeled Dissociation, Anxiety, Depression, Post-Sexual Abuse Trauma, Hypothesized and Sleep Disturbance. CSA has found to be associated with elevated scores on all of these subscales (Briere, Evans, Runtz & Wall, 1988). In addition, clinical samples report higher TSC scores than those based on non-clinical samples (Elliott & Briere, 1991). Seven items were added to the TCS-33 to create the TSC-40, which lead to the development of six subscales: Dissociation, Anxiety, Depression, Sexual Abuse Trauma Index, Sexual Problems and Sleep Disturbance. The TSC-40 was administered to a national stratified sample of professional women (*n* = 2,833; Elliott & Briere, 1990; *n* = 2,963; Elliott & Briere, 1992). Analyses revealed an average subscale alpha of .69 and alpha of .90 for the total measure. Thirty-six of the items discriminated between CSA victims and non-abused subjects in their non-clinical sample (Elliott & Briere, 1990). As expected, TSC-40 total scores correlate highly with the TCS-33 scores (r = .99; Elliott & Briere, 1991) and repeatedly discriminate between CSA survivors and non-survivors as indexed by verbal self-report and by therapist reports (Whiffen, Benazon & Bradshaw, 1997); Elliott & Briere, 1990; Elliott & Briere, 1992). The Sexual Abuse Trauma Index (Elliott & Briere, 1990) and Dissociation subscales (Elliott & Briere, 1992) were found to be the most predictive of sexual abuse history and is according to Elliott & Briere (1990) essentially equivalent to total TSC scores in terms of responsiveness to a history of CSA. In clinical and non-clinical samples, some abuse characteristics correlated significantly with total and subscale scores, and with the duration and frequency with which the abuse occurred (Elliott & Briere, 1991). Convergent validity of the TSC-40 has been established in its relationship to other measure of CSA and Trauma

promise and would profit from further validation.

(Brandyberry & MacNair-Semands, 1998; Gold & Cardena, 1998).

Despite its seemingly strong psychometric properties, the TSC-40 suffers some pitfalls. In one study, Whiffen, Benazon & Bradshaw (1997) found that the TSC-40 was no better at discriminating CSA than the SCL-90-R, a generic measure of symptomatology. In fact, for each measure (the SCL-90-R and TSC-40) the best predictors of CSA were the Anxiety

5. The Trauma Symptom Checklist TSC-40

The examination of the factor structure of the CAT revealed three factors; negativity of the home environment, sexual abuse; and punishment which accounted for 38% of the overall variance. Test-retest reliability of the subscales proved moderate with measures ranging from r = .71 to r = .91. However, males and females responded with significant differences on the sexual abuse subscale with females scoring twice as high which may reflect an over sensitivity to female responses. Furthermore, Kent & Waller (1998) reported that the intercorrelations between the scores on the subscales were strong with exception to the associations between abuse score and other scales. These correlations suggest that the CAT's subscales are not measuring entirely different constructs. Despite the CAT establishment of good reliability and other supporting psychometric data, to the author's knowledge, no studies have established criterion-related validity or have established its relationship to other abuse measures. Finally, the questionnaire items are relatively face-valid and susceptible to malingering.

4. The Childhood Trauma Questionnaire (CTQ)

The primary objective of the CTQ was to design an instrument which would allow for a comparison for both clinical and research purposes of a broad range of abuse and neglect experiences (Bernstein, Fink, Handelsman, Foote, Lovejoy, Wenzel, Sapareto & Ruggiero, 1994). The CTQ is a 70-item self-report questionnaire constructed from the author's extensive review of the literature which assess retrospectively childhood experience of abuse and neglect including aspects of their childhood rearing environment. Items are rated using a 5-point Likert-type scale.

The CTQ test-retest reliability was found to be .88 after an average interval of 3.6 months. A factor analysis of the CTQ yielded four factors accounting for 47.6% variance which were labeled physical and emotional abuse, emotional neglect, sexual abuse, and physical abuse. Convergent validity was assessed by correlating the CTQ with the Childhood Trauma Inventory (CTI) (Bernstein et al., 1994). After controlling for the effects of general maltreatment, the relation between the two sets of abuse rating was highly specific: CTQ sexual abuse was associated only with CTI sexual abuse and CTQ physical and emotional abuse was associated only with CTI physical abuse.

The CTQ has also shown good discriminent validity with measures of verbal intelligence and social desirability (r < .10 for all correlations; Bernstein et al., 1994) and convergent validity with the Traumatic Events Questionnaire – Adolescents (Weinzer & Lipshitz, 1997); the Traumatic History Screen (Allen, Huntoon & Evans, 1999); and other instruments designed to identify childhood maltreatment (Lipshitz, Bernstein, Wagner, & Southwick, 1999).

In addition, Bernstein, Ahluvalia, Pogge & Hanedelsman (1997) reported that CTQ cut-off scores have been devised for identifying cases of childhood abuse and neglect. In an adolescent psychiatric sample, a cut-off score of 9 on the Sexual Abuse factor had a sensitivity of .86 and specificity of .76 when therapists' ratings of sexual abuse were used as a criterion measure. Similarly, when using therapist's ratings of physical abuse as a criterion measure, a cut-off score of 12 on the physical abuse factor had a sensitivity of .82 and as specificity of .73.

Despite the widespread use of the CTQ and its supporting psychometric data, critics have pointed out that the original validation study examining convergence between the CTQ and CTI has shortcomings. For example, Bursten (1995) highlights the relative unimportance of the correlation between two forms of almost identical abuse measures and points out that these correlations tell us little of the questionnaire's ability to discriminate abuse victims. Bursten also criticized Bernstein et al. (1994) study for the author's use of the CIQ in their validation study because it was experimental and undergoing validation. Lastly, Bursten notes that aside from adolescent psychiatric patients, no normative data exists for the CTQ that would allow for its use in a clinical setting. Nevertheless, the CTQ shows considerable promise and would profit from further validation.

5. The Trauma Symptom Checklist TSC-40

8 Sexual Abuse – Breaking the Silence

The examination of the factor structure of the CAT revealed three factors; negativity of the home environment, sexual abuse; and punishment which accounted for 38% of the overall variance. Test-retest reliability of the subscales proved moderate with measures ranging from r = .71 to r = .91. However, males and females responded with significant differences on the sexual abuse subscale with females scoring twice as high which may reflect an over sensitivity to female responses. Furthermore, Kent & Waller (1998) reported that the intercorrelations between the scores on the subscales were strong with exception to the associations between abuse score and other scales. These correlations suggest that the CAT's subscales are not measuring entirely different constructs. Despite the CAT establishment of good reliability and other supporting psychometric data, to the author's knowledge, no studies have established criterion-related validity or have established its relationship to other abuse measures. Finally, the questionnaire items are relatively face-valid and

The primary objective of the CTQ was to design an instrument which would allow for a comparison for both clinical and research purposes of a broad range of abuse and neglect experiences (Bernstein, Fink, Handelsman, Foote, Lovejoy, Wenzel, Sapareto & Ruggiero, 1994). The CTQ is a 70-item self-report questionnaire constructed from the author's extensive review of the literature which assess retrospectively childhood experience of abuse and neglect including aspects of their childhood rearing environment. Items are rated

The CTQ test-retest reliability was found to be .88 after an average interval of 3.6 months. A factor analysis of the CTQ yielded four factors accounting for 47.6% variance which were labeled physical and emotional abuse, emotional neglect, sexual abuse, and physical abuse. Convergent validity was assessed by correlating the CTQ with the Childhood Trauma Inventory (CTI) (Bernstein et al., 1994). After controlling for the effects of general maltreatment, the relation between the two sets of abuse rating was highly specific: CTQ sexual abuse was associated only with CTI sexual abuse and CTQ physical and emotional

The CTQ has also shown good discriminent validity with measures of verbal intelligence and social desirability (r < .10 for all correlations; Bernstein et al., 1994) and convergent validity with the Traumatic Events Questionnaire – Adolescents (Weinzer & Lipshitz, 1997); the Traumatic History Screen (Allen, Huntoon & Evans, 1999); and other instruments designed to

In addition, Bernstein, Ahluvalia, Pogge & Hanedelsman (1997) reported that CTQ cut-off scores have been devised for identifying cases of childhood abuse and neglect. In an adolescent psychiatric sample, a cut-off score of 9 on the Sexual Abuse factor had a sensitivity of .86 and specificity of .76 when therapists' ratings of sexual abuse were used as a criterion measure. Similarly, when using therapist's ratings of physical abuse as a criterion measure, a cut-off score of 12 on the physical abuse factor had a sensitivity of .82 and as

Despite the widespread use of the CTQ and its supporting psychometric data, critics have pointed out that the original validation study examining convergence between the CTQ and CTI has shortcomings. For example, Bursten (1995) highlights the relative unimportance of

identify childhood maltreatment (Lipshitz, Bernstein, Wagner, & Southwick, 1999).

susceptible to malingering.

using a 5-point Likert-type scale.

specificity of .73.

4. The Childhood Trauma Questionnaire (CTQ)

abuse was associated only with CTI physical abuse.

The TSC-40 appears to be the most popular childhood abuse questionnaire available (Castelda, 2003) perhaps in part because of extensive research history and strong psychometric support. TSC-40 was developed from the Trauma symptom Checklist – TSC-33 (Briere & Runtz, 1989) which was developed to assess the impact of trauma especially the long-term effects of CSA. The emphasis was on gaining a more complete understanding of the precise patterns of symptomatology associated with a history of CSA (Gold, Milan, Mayall & Johnson, 1994). Participants respond to items by rating how often they had experienced each in the last two months (0 = never, 4 = very often. In a sample of woman at a crisis intervention clinic, TSC-33 was found to be internally consistent, alpha = .89 capable of correctly identifying 79% of self-reported CSA survivors. Twenty-three of these items discriminated between abused and non-abused participants. The TSC-33 consists of five subscales labeled Dissociation, Anxiety, Depression, Post-Sexual Abuse Trauma, Hypothesized and Sleep Disturbance. CSA has found to be associated with elevated scores on all of these subscales (Briere, Evans, Runtz & Wall, 1988). In addition, clinical samples report higher TSC scores than those based on non-clinical samples (Elliott & Briere, 1991).

Seven items were added to the TCS-33 to create the TSC-40, which lead to the development of six subscales: Dissociation, Anxiety, Depression, Sexual Abuse Trauma Index, Sexual Problems and Sleep Disturbance. The TSC-40 was administered to a national stratified sample of professional women (*n* = 2,833; Elliott & Briere, 1990; *n* = 2,963; Elliott & Briere, 1992). Analyses revealed an average subscale alpha of .69 and alpha of .90 for the total measure. Thirty-six of the items discriminated between CSA victims and non-abused subjects in their non-clinical sample (Elliott & Briere, 1990). As expected, TSC-40 total scores correlate highly with the TCS-33 scores (r = .99; Elliott & Briere, 1991) and repeatedly discriminate between CSA survivors and non-survivors as indexed by verbal self-report and by therapist reports (Whiffen, Benazon & Bradshaw, 1997); Elliott & Briere, 1990; Elliott & Briere, 1992). The Sexual Abuse Trauma Index (Elliott & Briere, 1990) and Dissociation subscales (Elliott & Briere, 1992) were found to be the most predictive of sexual abuse history and is according to Elliott & Briere (1990) essentially equivalent to total TSC scores in terms of responsiveness to a history of CSA. In clinical and non-clinical samples, some abuse characteristics correlated significantly with total and subscale scores, and with the duration and frequency with which the abuse occurred (Elliott & Briere, 1991). Convergent validity of the TSC-40 has been established in its relationship to other measure of CSA and Trauma (Brandyberry & MacNair-Semands, 1998; Gold & Cardena, 1998).

Despite its seemingly strong psychometric properties, the TSC-40 suffers some pitfalls. In one study, Whiffen, Benazon & Bradshaw (1997) found that the TSC-40 was no better at discriminating CSA than the SCL-90-R, a generic measure of symptomatology. In fact, for each measure (the SCL-90-R and TSC-40) the best predictors of CSA were the Anxiety

A Review of Childhood Abuse Questionnaires and Suggested Treatment Approaches 11

found that total TSI scores have been found to be predictive of Borderline Personality

In a regression analysis it was shown that CSA was strongly associated with all ten subscales. CPA was the next best predictor of TSI scores, predicting all scales but the two involving sexual difficulties (Briere et al., 1994). Furthermore, the TSI show promise as a means for identifying persons motivated to exaggerate or fabricate symptomatology as well as malingering responses styles among those claiming psychological damage resulting from

Although the focus on symptom correlates of CSA has proven to be a valuable approach in the attempt to assess the presence of a CSA history, it does not directly address the

Over the past fifteen years, our laboratory has been devoted to the development of an alternative non-symptom childhood abuse questionnaire designed to increase the sensitivity of the measure's ability to detect a history of CSA. The goal was to develop a time efficient non-face-valid screen via the selection of questions that primarily query sequelae and correlates of CA. The items selected for testing were primarily selected from the present author's research data base of over 5000 treatment transcripts of adult delayed memory recall CA patients (Levis, 1995; Levis & Brewer, 2001). For example, true/false items such as "I have not had a happy childhood"; "at times I have had the feeling that I have fallen into a

The first attempt to implement the above strategy consisted of a 103 true-false item screen, referred to as the Sexual Abuse Questionnaire (SAQ-I). To insure the functional utility of the SAQ, the initial validation was conducted with non-clinical undergraduates population who self-reported a history of CSA prior to the age of 15 on Part II of the SAQ which consisted of a number of open-ended questions designed to establish the presence of a history of CSA. Alexander (1993) provided support for the initial use of a non-clinical research population. He reported that the symptomatology of his non-clinical population of women who selfreported a history of incestuous abuse was comparable to that reported by a clinical

Lock, Levis, and Rourke (2005) provided the initial evaluation of the SAQ-I. Study I (N=548) established the SAQ's test-retest reliability, internal-consistency and convergent validity by comparing it to Eliot & Briere's (1992) Trauma Symptom Checklist (TSC-40) and to the MMPI subscale for post-traumatic stress (Keane, Malloy, & Fairbank, 1984). SAQ-I demonstrated good reliability (r=.84) and adequate internal consistency (KR-20=.89). Convergent validity was also established with the TSC-46 and Keane's MMPI subscale. After removing similar items, the SAQ correlated moderately with the TSC-40 (r=.63) and the Kean's PTSD subscale (r=.68). Thirty-nine percent of the SAQ-I items discriminated between abused and non-abused groups, compared to 35% of the TSC-40 items. Using a cutting score of 41 for males, the SAQ-I correctly classified 68% of the sample (63% of the abused participants and 69% of the non-abused sample. For women, a cutting score of 42 was established which correctly classified 67% of the sample (64% of the abused participants and 68% of the non-abused sample). The SAQ-I was able to discriminate between abused and non-abused females and males while the TSC-40 failed to discriminate between the

methodical concerns raised above which inhibit the detection of a CSA history.

black hole or cloud"; and, "I am much lonelier than most people".

Disorder as well as PTSD.

population.

a traumatic experience (Edens, Otto, Dwyer, 1998).

7. The Binghamton Childhood Abuse Screen (BCAS)

subscale and the PSAT, respectively. They point out that the PSAT subscale, which is made of only a few generally worded items, includes items relating to nightmares and fears of men, suggesting that both subscales tap PTSD symptoms. They suggest the existing TSC-40 PSAT subscale items be put into more specific terms and that more items be added to the TSC-40 PSAT subscale to allow for a better characterization of CSA survivors (Whiffen, Benazon & Bradshaw, 1997). Other studies have shown that the TSC-40 not only discriminates between those with a history of CSA but also adult-experienced sexual abuse (Gold, Milan, Mayall & Johnson, 1994). Similar to the above findings, these results suggest that the TSC-40 is not entirely sensitive to the sequelae of CSA; rather it appears to tap into the trauma associated with sexual abuse in general.

The validity of the TSC-40 may also be improved by addressing some psychometric problems. One such issue is the inclusion of items in more than one subscale, thereby inflating intercorrelations. Whiffen, Benazon & Bradshaw (1997) suggest that by confirming the hypothesized structure through factor analysis, items could be independently assigned to appropriate subscales, and therefore purifying the subscales and improving the validity of the instrument.

Other criticisms of the TSC-40 have focused on the standardization sample which consists exclusively of professional women, limiting the generalizability of scores to men, minority women, and women with annual incomes under \$30,000 (Lock, Levis & Rourke, 2003). Further deficiencies include: a lack of test-retest reliability estimates, overlapping standard deviations of abused and non-abused individuals which limits the interpretability of total scores; and a lack of external corroboration for the self-reported CSA criterion (Lock, Levis & Rourke, 2003).

6. The Trauma Symptom Inventory (TSI)

The TSI, developed by Briere (1995) is a 100-item, structure, self-report inventory designed for general clinical use in the assessment of the psychological sequelae of potentially traumatic events, including physical violence and sexual assault (Briere & Elliot, 1997). Its construction was in response to the paucity of standardized, clinically useful measures of posttraumatic symptomatology (Briere, Elliott, Harris & Cotman, 1995). Items were developed from an expansion of the TSC-40 and from clinical hypotheses. The test items are rated according to the frequency of symptoms over the prior six months on a four-point scale ranging from 0 ("never") to 3 ("often"). The TSI uses ten clinical scales and three validity scales. The clinical scales include: Anxious Arousal, Depression, Anger/Irritability, Intrusive Experiences, Defensive Avoidance, Dissociation, Sexual Concerns, Dysfunctional Sexual Behavior, Impaired Self-Reference, and Tension-Reduction Behavior. The validity scales include: Response Level, Measuring General Under-endorsement, Atypical Response, Evaluating Over-endorsement, and Inconsistent Response which measures inconsistent responses between similar TSI items-pairs (Briere, 1995).

The TSI was standardized using a random sample of 828 people representative of the United States and using 3,659 military recruits. Norms are available for four combinations of sex and age (Briere, 1995). The clinical scales of the TSI are relatively consistent (α = .84 - .87) and exhibit reasonable predictive and incremental validity (Briere, 1995). The TSI total scores also demonstrate good convergent validity with independently assessed PTSD status with a specificity of .92 and sensitivity of .91 (Astin, Lawrence & Foy, 1993). Briere (1995) found that total TSI scores have been found to be predictive of Borderline Personality Disorder as well as PTSD.

In a regression analysis it was shown that CSA was strongly associated with all ten subscales. CPA was the next best predictor of TSI scores, predicting all scales but the two involving sexual difficulties (Briere et al., 1994). Furthermore, the TSI show promise as a means for identifying persons motivated to exaggerate or fabricate symptomatology as well as malingering responses styles among those claiming psychological damage resulting from a traumatic experience (Edens, Otto, Dwyer, 1998).

7. The Binghamton Childhood Abuse Screen (BCAS)

10 Sexual Abuse – Breaking the Silence

subscale and the PSAT, respectively. They point out that the PSAT subscale, which is made of only a few generally worded items, includes items relating to nightmares and fears of men, suggesting that both subscales tap PTSD symptoms. They suggest the existing TSC-40 PSAT subscale items be put into more specific terms and that more items be added to the TSC-40 PSAT subscale to allow for a better characterization of CSA survivors (Whiffen, Benazon & Bradshaw, 1997). Other studies have shown that the TSC-40 not only discriminates between those with a history of CSA but also adult-experienced sexual abuse (Gold, Milan, Mayall & Johnson, 1994). Similar to the above findings, these results suggest that the TSC-40 is not entirely sensitive to the sequelae of CSA; rather it appears to tap into

The validity of the TSC-40 may also be improved by addressing some psychometric problems. One such issue is the inclusion of items in more than one subscale, thereby inflating intercorrelations. Whiffen, Benazon & Bradshaw (1997) suggest that by confirming the hypothesized structure through factor analysis, items could be independently assigned to appropriate subscales, and therefore purifying the subscales and improving the validity

Other criticisms of the TSC-40 have focused on the standardization sample which consists exclusively of professional women, limiting the generalizability of scores to men, minority women, and women with annual incomes under \$30,000 (Lock, Levis & Rourke, 2003). Further deficiencies include: a lack of test-retest reliability estimates, overlapping standard deviations of abused and non-abused individuals which limits the interpretability of total scores; and a lack of external corroboration for the self-reported CSA criterion (Lock, Levis

The TSI, developed by Briere (1995) is a 100-item, structure, self-report inventory designed for general clinical use in the assessment of the psychological sequelae of potentially traumatic events, including physical violence and sexual assault (Briere & Elliot, 1997). Its construction was in response to the paucity of standardized, clinically useful measures of posttraumatic symptomatology (Briere, Elliott, Harris & Cotman, 1995). Items were developed from an expansion of the TSC-40 and from clinical hypotheses. The test items are rated according to the frequency of symptoms over the prior six months on a four-point scale ranging from 0 ("never") to 3 ("often"). The TSI uses ten clinical scales and three validity scales. The clinical scales include: Anxious Arousal, Depression, Anger/Irritability, Intrusive Experiences, Defensive Avoidance, Dissociation, Sexual Concerns, Dysfunctional Sexual Behavior, Impaired Self-Reference, and Tension-Reduction Behavior. The validity scales include: Response Level, Measuring General Under-endorsement, Atypical Response, Evaluating Over-endorsement, and Inconsistent Response which measures

The TSI was standardized using a random sample of 828 people representative of the United States and using 3,659 military recruits. Norms are available for four combinations of sex and age (Briere, 1995). The clinical scales of the TSI are relatively consistent (α = .84 - .87) and exhibit reasonable predictive and incremental validity (Briere, 1995). The TSI total scores also demonstrate good convergent validity with independently assessed PTSD status with a specificity of .92 and sensitivity of .91 (Astin, Lawrence & Foy, 1993). Briere (1995)

inconsistent responses between similar TSI items-pairs (Briere, 1995).

the trauma associated with sexual abuse in general.

6. The Trauma Symptom Inventory (TSI)

of the instrument.

& Rourke, 2003).

Although the focus on symptom correlates of CSA has proven to be a valuable approach in the attempt to assess the presence of a CSA history, it does not directly address the methodical concerns raised above which inhibit the detection of a CSA history.

Over the past fifteen years, our laboratory has been devoted to the development of an alternative non-symptom childhood abuse questionnaire designed to increase the sensitivity of the measure's ability to detect a history of CSA. The goal was to develop a time efficient non-face-valid screen via the selection of questions that primarily query sequelae and correlates of CA. The items selected for testing were primarily selected from the present author's research data base of over 5000 treatment transcripts of adult delayed memory recall CA patients (Levis, 1995; Levis & Brewer, 2001). For example, true/false items such as "I have not had a happy childhood"; "at times I have had the feeling that I have fallen into a black hole or cloud"; and, "I am much lonelier than most people".

The first attempt to implement the above strategy consisted of a 103 true-false item screen, referred to as the Sexual Abuse Questionnaire (SAQ-I). To insure the functional utility of the SAQ, the initial validation was conducted with non-clinical undergraduates population who self-reported a history of CSA prior to the age of 15 on Part II of the SAQ which consisted of a number of open-ended questions designed to establish the presence of a history of CSA. Alexander (1993) provided support for the initial use of a non-clinical research population. He reported that the symptomatology of his non-clinical population of women who selfreported a history of incestuous abuse was comparable to that reported by a clinical population.

Lock, Levis, and Rourke (2005) provided the initial evaluation of the SAQ-I. Study I (N=548) established the SAQ's test-retest reliability, internal-consistency and convergent validity by comparing it to Eliot & Briere's (1992) Trauma Symptom Checklist (TSC-40) and to the MMPI subscale for post-traumatic stress (Keane, Malloy, & Fairbank, 1984). SAQ-I demonstrated good reliability (r=.84) and adequate internal consistency (KR-20=.89). Convergent validity was also established with the TSC-46 and Keane's MMPI subscale. After removing similar items, the SAQ correlated moderately with the TSC-40 (r=.63) and the Kean's PTSD subscale (r=.68). Thirty-nine percent of the SAQ-I items discriminated between abused and non-abused groups, compared to 35% of the TSC-40 items. Using a cutting score of 41 for males, the SAQ-I correctly classified 68% of the sample (63% of the abused participants and 69% of the non-abused sample. For women, a cutting score of 42 was established which correctly classified 67% of the sample (64% of the abused participants and 68% of the non-abused sample). The SAQ-I was able to discriminate between abused and non-abused females and males while the TSC-40 failed to discriminate between the

A Review of Childhood Abuse Questionnaires and Suggested Treatment Approaches 13

points to the need to assess the possibility that a history may exist when developing a treatment plan. The use of one of the reviewed abuse assessment inventories should not only facilitate this assessment goal but it also helps communicate to the patient the willingness of the therapists to deal with this topic. It is surprising to me that a large number of patients who I have seen have reported their previous therapist never raised the issue of childhood abuse. It is unfortunate that too many of our mental health workers are still in a state of denial regarding the impact an abuse history can have on the client's behavior. If no support for a history of abuse is found, this possibly should remain open given the finding

If a history of childhood abuse is self-reported, the issue of treatment selection becomes paramount. Unfortunately, a review of all the existing treatment approaches would require another chapter. Nevertheless, I will briefly discuss two approaches that have yielded some

The first technique to be discussed is the technique of Cognitive Behavior Therapy (CBT). The technique of CBT was designed to be used to treat a wide variety of clinical symptoms including cases with a history of childhood abuse. CBT has been subjected to experimental

CBT has been defined as an intervention whose core elements include the recipient establishing links between their thoughts, feelings and actions and targeting symptoms correcting misperceptions, irrational beliefs, and reasoning biases related to these target symptoms, involving monitoring of one's own thoughts, feelings and behaviors with respect to the symptom; and/or the promotion of alternative ways of coping with target symptoms

Although CBT is a commonly used approach, the variety of techniques used to create cognitive restriction varies from one study to the other, frequently lacking operational specificity. Given the number of strategies and approaches used to change cognitive behavior, it appears to me to be similar to the insight approaches adapted in the 1940's and 1950's, (Alexander, 1965). It can be argued that the key agent of change in CBT is the approach ability to elicit emotional affects which in turn undergoes an extinction process that in turn results in a cognitive restructuring. This hypothesis leads to the discussion of the

The technique of prolonged CS exposure was first developed by Thomas G. Stampfl in the late 1950's. Despite considerable pressure to publish, he refused to publish until two experimental outcome studies were conducted to support his clinical finding. They were provided by Hogan (1966) who used a psychiatric hospitalized population, and Levis & Carrera (1967) who used an outpatient population. He then published his first article (Stampfl & Levis, 1967). He labeled his new response prevention approach, Implosive Therapy (IT). He first used an in vivo extinction approach but changed to an imagery approach due to the ability to introduce hypothesized conditioned stimulus (CS) that do not lend themselves readily to in vivo exposure (e.g., the fear of bodily injury). Stampfl adapted

validations (see Lynch, Lowe & McKenna, 2010) for a review of supporting data.

of delayed memory recall.

experimental support.

**5. Cognitive Behavioral Therapy (CBT)** 

(Lynch, Lowe & McKenna, 2010).

second approach to be discussed.

**6. The use of Prolonged CS Exposure (PE)** 

reported abused and non-abused males. An exploratory factor analysis that utilized a maximum likelihood extraction and orthogonal rotation suggested a one-dimensional structure. Based on an item analysis of SAQ-I, 78 items were retained for the second version of the SAQ.

Study II (N=533), in addition to examining the psychometric properties of SAQ-II, also assessed the discriminant validity using the Beck Depressive Inventory (Beck, Steer, & Garbin, 1988) and the trait component of the State-Trait Anxiety Inventory (Spielberger, 1983). The SAQ-II correctly identified 86% of the male sample (73% of the abused participants and 86% of the non-abused participants). For women, 73% of the sample was correctly identified (64% of the abused participants and 74% of the non-abused participants). The SAQ-II was found to be a better post-dictor of abuse than the other two measures combined. The SAQ-II also yielded convergent validity with the TSC-40 (r=.67) and Keane's PTSSD subscale (r=.66). An exploratory factor analysis again suggested a one dimensional structure (see Lock, Levis, & Rourke, 2005). Item analysis of the SAQ-II resulted in the development of SAQ-III, a 45 item questionnaire. The SAQ-III was found to correctly identify a participant abuse history. Eighty-nine percent of the sample tested (83% of the male participants and 91% of the female participants were correctly identified (see Krantweiss, 2001, 2004).

Castelda (2006) provided an examination of the relationship between the SAQ-III and the complex post-traumatic disorder (Lenzenweger, Loranges, Korfine, & Neff, 1997). Results showed that those scoring high on the SAQ reported more incidence of childhood abuse, greater levels of PTSD and Axis II psychopathology, and displayed smaller heart rate changes during an auditory startle test than those scoring low on the SAQ.

Evidence has also been found which suggested the SAQ-III is also capable of identifying survivors of childhood physical abuse (see Krantweiss, 2004). Castelda (2003) confirmed Krantweiss's findings along with providing support for SAQ-III's ability to identify survivors of combined CSA and CPA histories. These findings are reconfirmed in Experiment I of Castelda, Levis, Rourke, and Coleman (2007) publication using a large sample of participant (N=3,505). Experiment II of this study evaluated the effectiveness of a newly developed 36 item abuse screen which was renamed the Binghamton Childhood Abuse Screen (BCAS). An ROC survey analysis was provided for determining the sensitivity and specificity indices for all possible BCAS cut-off scores across all abuse types, enabling examiners to choose cut-off scores that are suitable for their purposes. Additional criterion validity for the BCAS was established via the use of a modified Stroop task by Coleman, Rourke, and Levis (2008).

Finally, the issue of malingering regarding the BCAS has been evaluated and support for the BCAS claim that it represent a non-face valid screen, has been obtained (Levis, Rourke, Bovier, Coleman, Heron, Castelda, and Esch, 2011). Although the BCAS has received sufficient support for its use with an undergraduate college population, it has yet to be validated on an out-patient population. However, some support for its use with an inpatient population has been recently obtained (Rowland, Ocelnik, Berryman, and Levis, 2011). Although the BCAS shows promise, it still remains a work in progress.

The previous review of the prevalence rates of childhood abuse in our society and the existing established link between childhood abuse and clinical symptomatology clearly points to the need to assess the possibility that a history may exist when developing a treatment plan. The use of one of the reviewed abuse assessment inventories should not only facilitate this assessment goal but it also helps communicate to the patient the willingness of the therapists to deal with this topic. It is surprising to me that a large number of patients who I have seen have reported their previous therapist never raised the issue of childhood abuse. It is unfortunate that too many of our mental health workers are still in a state of denial regarding the impact an abuse history can have on the client's behavior. If no support for a history of abuse is found, this possibly should remain open given the finding of delayed memory recall.

If a history of childhood abuse is self-reported, the issue of treatment selection becomes paramount. Unfortunately, a review of all the existing treatment approaches would require another chapter. Nevertheless, I will briefly discuss two approaches that have yielded some experimental support.
