The Danger Assessment Instrument (DA), in its current form, is a 20-item actuarial test designed to assess the risk of serious or lethal intimate partner violence. It is intended for use with adult women who have suffered physical abuse at the hands of men who are their current or former intimate partners. Although originally developed to assist in safety planning conducted by people delivering services to victims, the DA more recently has been used by some law enforcement agencies to help manage the risks posed by perpetrators. Systematic review of the DA is complicated by the fact that it has been used in several different forms for a variety of different purposes and by the lack of a formal test manual.
Description and Development
Development of the DA began in the early 1980s. The DA differs from other tests designed to assess the risk of intimate partner violence—such as the Spousal Assault Risk Assessment Guide (SARA) and the Domestic Violence Screening Instrument (DVSI)—in that its development focused specifically on risk of serious or lethal violence. In its original form, it comprised 15 risk factors that were identified in retrospective studies of intimate partner homicide—cases in which battered women killed or seriously injured their abusive partners or in which battered women were killed or seriously injured by their abuser partners. In 2004, the DA item pool was revised and expanded to 20 items, based on the findings of a multisite study that compared risk factors for life-threatening (lethal or near lethal) versus less serious intimate partner violence. Some of the items reflect the nature or severity of intimate partner violence in the victim’s relationship with the perpetrator, such as a history of threats to kill, forced sex, or strangulation; some reflect characteristics of the victim, such as whether she has children from a previous relationship or has a history of suicidal threats or attempts; and others reflect characteristics of the perpetrator, such as whether he has a history of problems with employment or substance use.
The DA can be completed independently or in collaboration with a service provider. Administration of the DA begins with a calendar assessment, in which the victim reviews a calendar to determine the nature of frequency of intimate partner violence experienced by the victim in the previous year. The victim identifies the approximate dates of any abuse and rates the seriousness of each incident using a 5-point scale (1 = slapping, pushing; no injuries and/or lasting pain to 5 = use of weapon; wounds from weapon). This history taking gathers information that is useful when rating the 20 risk factors, but it also is intended to decrease the extent to which victims minimize the intimate partner violence that they have experienced.
Next, victims are asked to rate the presence of the 20 risk factors on a 2-point scale (0 = no, 1 = yes). Some are rated on the basis of lifetime presence, whereas others are rated on their presence in the past year. Items ratings are then summed using a simple unit-weighting procedure to yield total scores that range from 0 to 20; alternatively, a more complex dif-ferential weighting procedure can be used that yields total scores ranging from -3 to 37. Total scores can be classified into four categories that reflect the risk of life-threatening violence: <7 = variable danger, 8 to 13 = increased danger, 14 to 17 = severe danger, and >18 = extreme danger. The risk categories are associated with suggested intervention strategies and directions regarding what should be communicated to the victim. For example, if a victim’s DA score falls in the severe danger category, the service provider is advised to inform the victim that she is in severe danger, engage in assertive safety planning with her, and recommend a high level of supervision for the perpetrator.
The psychometric properties of different versions of the DA have been evaluated only to a limit extent and solely within the framework of classical test theory.
Evaluations of structural reliability have reported Cronbach’s alpha averaging about .75 to .80, and evaluations of short-term test-retest reliability have reported correlations averaging about .90. Given that the DA is an actuarial test of violence risk, these findings are actually somewhat disappointing: First, they indicate that the risk factors included in the test are at least moderately correlated, which suggests that they are likely to be substantially redundant as predictors of intimate partner violence. Second, they indicate that the DA is apparently insensitive to short-term changes or fluctuations in violence risk.
Little or no information is available concerning the interrater reliability of the DA—that is, agreement between victims and evaluators or agreement among evaluators with respect to the item or total scores.
There have been no studies evaluating the psychometric properties of the DA within the framework of modern test theory.
Some support for the validity of the DA comes from retrospective studies that found a significant association between total scores and measures of the seriousness of past intimate partner violence. Some research has attempted to determine whether the DA can dis-criminate between victims of lethal (or life threatening) versus nonlethal intimate partner violence, with disappointing results; but this may be because the DA must be scored on the basis of information provided by collateral informants (e.g., relatives, friends) when victims are deceased, possibly resulting in decreased validity of test scores.
There is also some research supporting the predictive validity of the DA with respect to intimate partner violence. First, indirect evidence comes from studies that found moderate to high correlations, typically between .55 and .75, between DA total scores and scores on other measures related to risk of intimate partner violence, such as the Conflict Tactics Scale and the Index of Spouse Abuse. Second, direct evidence comes from prospective studies that have found a moderate association between DA total scores and repeated intimate partner violence. Although there have been few direct comparisons, the predictive validity of the DA appears to be about the same as that of other procedures for assessing risk of intimate partner violence. To date, there has been no investigation of the predictive validity of the DA specifically with respect to life-threatening violence, the purpose for which it was originally developed.
The DA can be useful as part of a comprehensive assessment of risk of intimate partner violence. It has two important strengths: It systematically gathers information from victims, who can provide a unique perspective on the history of violence in the relationship and on the perpetrator’s background and psychosocial adjustment, and it considers victim vulnerability factors that are relevant to safety planning.
The DA also has some important limitations. There is no formal manual to guide proper administration, scoring, and interpretation or that provides the technical information necessary to undertake a comprehensive review of the test. There is a lack of information concerning the interrater reliability of DA item and total scores. There is a lack of systematic research on the predictive validity of the DA with respect to intimate partner violence in general and no research with respect to life-threatening intimate partner violence. At the present time, then, it may be best to use the DA as structured professional guidelines for risk assessment rather than as a quantitative or actuarial test.
- Campbell, J. C. (1995). Assessing dangerousness. Thousand Oaks, CA: Sage.
- Campbell, J. C., Webster, D., Koziol-McLain, J., Block, C., Campbell, D., Curry, M., et al. (2003). Risk factors for femicide in abusive relationships: Results from a multisite case control study. American Journal of Public Health, 93, 1089-1097.