Screening Measures for Domestic
Violence, Sexual Assault, and Physical Assault
Violent crime, particularly that type of crime associated with interpersonal, psychological, or cultural stigma (e.g., rape) is not readily reported by all victims. Indeed, in order to report to an investigator that a particular type of crime has occurred, a victim must:
While straightforward, these factors must not be taken for granted. For example, many respondents do not label aggravated assault or rape as such when the perpetrator is a relative or spouse, or when there was only limited force or threat of force used, or when the psychological effects of such a label are too distressing. Hence, again asking "Have you ever been raped?" will result in overly conservative prevalence estimates (Essock-Vitale & McGuire, 1985; Gordon & Riger, 1985; Koss, 1993). Further, many victims are exceedingly reluctant to disclose their victimization experiences. Reasons for willful non-disclosure include: (1) fear of retribution by an assailant, particularly if the assailant is known or proximate to victim, (2) fear of stigma attached to being a victim of a particular type of crime (e.g., rape, domestic violence), (3) fear of being blamed, (4) history of negative outcomes following previous disclosure (e.g., court involvement leading to acquittal), (5) lack of encouragement to discuss victimization, and (6) fear of psychological consequences of disclosure (e.g., depression, anxiety upon re-visiting the event) (Kilpatrick, 1983; Koss, 1993; Resnick et al., 1996). It should be obvious that investigators conducting prevalence studies must not assume that all victimization events will be specifically and easily reported. Unfortunately, this stipulation has not always been met.
Given that the above factors will combine to reduce the likelihood that a crime event will be reported, what procedural aspects of assessment are essential to maximize sensitivity? Two components appear crucial:
Because traumatic events such as violent crime are associated with extremely aversive emotional and cognitive states, it is important, both to respondent welfare and experimental integrity, to preface criminal victimization queries in such a way as to convey acceptance, empathy, normalization, and encouragement. After the preface statement orients a respondent contextually (e.g., any assault by any assailant), it is absolutely essential to employ behaviorally-specific detailed, closed-ended descriptions of trauma events under investigation. Early surveys employed "gateway" screening questions characterized by very limited behavioral specificity. If respondents endorsed the gateway question, further questions about assault followed. Unfortunately, gateway questions do not adequately orient respondents to the type of responses the examiner is seeking (i.e., they fail to state that one is interested in all assaults, not just those reported to police or perpetrated by strangers) and are extremely subject to individual respondentís subjective interpretation of queries (i.e., definitional variance) (Koss, 1993).
Assessment of assault-related traumatic events is currently achieved, with varying degrees of success, through standardized questionnaires and structured clinical interviews. As Resnick et al. (1996) noted, inquiries that produce dichotomous data regarding event occurrence/non-occurrence should be complemented by questions that generate qualitative data regarding event characteristics. In this manner, multi-variate relationships between trauma factors can be assessed and targeted by interventions.
Survey of Exposure to Community Violence (SCECV; Richters & Saltzman, 1990): This survey examines environmental and systems-level exposure to violence. Other relevant sections include arrest experiences, home break-ins, and specific weapons-related violence. For each type of violence, respondents rate direct experience of each type of violence on a 4 point scale. Recency of each type of incident is also recorded. These ratings yield a total continuous score of direct exposure to community violence. Although the measure was designed for use with adolescents it may also be used with adults.
The Revised Conflict Tactics Scales (CTS2): The CTS2 (Straus, Hamby, Boney-McCoy, & Sugarman, 1996), originally developed by Straus (1979) is a widely used (over 70,000 empirical studies have used it) and thoroughly evaluated (approximately 400 papers) measure of interpersonal aggression in married or cohabitating partners. Note that it is not a measure of attitudes toward violence, but rather, a measure of conflict resolution events that involve violence. The scales also measure psychological abusiveness and the use of negotiation and reasoning by either partner to reduce conflict Although the CTS has undergone numerous revisions in the past 15 years, its basic structure has remained the same. The most recent version contains several scales: reasoning/negotiation (6 items), psychological aggression (8 items), physical assault (12 items), sexual coercion (7 items), and a consequence (physical injury) (6 items) scale. The 39 items are rated on a 8-point frequency scale (never, once, twice, 3-5 times, 6-10 times, 11-20 times, and more than 20 times, not in the past year but it did happen before). Interpersonal problem-resolution behaviors range from benign (e.g., A...when you had a dispute have spouse discussed the issue calmly) to dangerous (e.g., "Has your spouse threatened you with a knife or gun). Each question is asked in terms of both respondentís and partnerís behavior. Reliability ranges from .79 to .95 and initial evidence of construct validity has been obtained (reliability and validity of the scale are well-established, and early factor analysis revealed constructs representing 1) verbal reasoning, 2) psychological abuse/aggression, 3) physical aggression, and 4) life-threatening violence (see Strauss, 1990 for a thorough psychometric summarization).
The PTSD module of the Structured Clinical Interview for DSM-III-R (SCID-R) and the newer version for DSM-IV (SCID-IV) contains a very brief preface describing possible PTSD Criterion A events, followed by a short trauma priming list. No behavioral definitions of assault are provided, and respondents are required to generate descriptions of any experienced trauma that are not on the list. Although the interviewer documents all traumatic events, only the subjectively defined "worst" trauma is considered in the diagnostic section. No attempt is made to formally identify or evaluate subjective aspects of trauma that relate to outcome, such as perceived risk of severe injury or death. Similarly, no specific prompts are provided to elicit relevant objective event characteristics, such as number of assailants, duration of assault, presence of a weapon, etc.. Overall, lack of behaviorally-defined close-ended assault questions, weaknesses of the preface statement, event documentation and prompting, and the extent to which objective and subjective parameters of assault are overlooked lessen the likelihood that SCID-based assessments of assault trauma will be of sufficient sensitivity to detect all instances of interpersonal aggression.
The Diagnostic Interview Schedule-III-Revised (DIS-III-R) (Robins, Helzer, Cottler, & Goldring, 1988) is a structured clinical interview similar in emphasis and format to the SCID. Specifically, a brief preface statement is followed by a listing of traumatic events intended to prime respondents to describe their own trauma history. As with the SCID, no behavioral descriptions of assault events are provided. In addition to the index question, respondents are prompted to report up to two additional "terrible" or "shocking" experiences they have had. In contrast to the SCID, the DIS-R qualitatively assesses whether injury resulted in response to the traumatic event, but does not specifically measure other objective or subjective aspects of assault.
Both the SCID and the DIS-R emphasize assessment of PTSD symptomatology over that of trauma per se. While achieving acceptable diagnostic reliability, both interviews fail to employ sufficiently sensitive or valid techniques to measure trauma. This is somewhat ironic and distressing in that no diagnosis of PTSD is possible without preliminary positive identification of a traumatic event. This is particularly problematic when considering that the overwhelming majority of treatment outcome and epidemiological studies, in addition to hundreds of clinic practices, employ these structured interviews to assess psychopathology. An effective solution to this sensitivity problem involves pairing these symptom-focused interviews with a trauma-focused assessment instrument. Such a hybrid has been developed by Kilpatrick et al. (1989) for the National Womens Study (NWS).
The NWS Event History-PTSD Module represents the synthesis of several trauma assessment devices refined through years of epidemiological research on physical and sexual assault by the National Crime Victims Research and Treatment Center at the Medical University of South Carolina. This interview comprehensively assesses lifetime occurrence of civilian crime, and includes a preface statement that provides contextual orientation to the trauma victim, along with accurate information regarding criterion A event prevalence. Moreover, the preface statement details the interviewerís interest in any assault event, not just those perpetrated by strangers or reported to police or individuals in a position of authority.
In addition to physical assault, the NWS Event History PTSD Module assesses
other typically stressful life events (e.g., automobile accidents, natural
disasters). Past-year occurrence of common life stressors (e.g., marital
discord, financial difficulty) are addressed in another section of the
interview. Sensitivity is further enhanced and popular stereotypical definitions
of trauma terminology avoided through use of behaviorally-specific closed-ended
questions about physical and sexual assault. Both objective and subjective
trauma-related data are obtained, including number of lifetime interpersonal
violence episodes, age at first incident of each trauma type, relationship
to perpetrator, personal and perpetrator use of substances preceding attack,
most serious incident of each trauma type, and perceptions of imminent
injury or death from the attack. Although all victimization events are
addressed, qualitative information about physical assault history is gathered
for only one event (as opposed to rape, for which information is gathered
in response to the first, most recent, and "worst" victimization). Further,
the measure is geared toward severe assault and may produce overly conservative
estimates of some forms of domestic violence. Convergent validity of the
instrument has been established, with estimated rates of sexual and physical
assault approximating those of other studies (e.g., Norris, 1992).