The Development of the Brief Spousal Assault Form for the Evaluation of Risk (B-SAFER): A Tool for Criminal Justice Professionals

Appendix B (continued)

Administration Procedure

It is imperative that all available sources of information are utilized when conducting a risk assessment. The B-SAFER should not be completed until a detailed assessment has been conducted; factors can be coded after all information has been collected and weighed. Ideally, the assessment will include:

We recommend that evaluators discuss in any written or oral reports the completeness of the information on which the risk assessment was based, as well as any limitations on that opinion due to missing and/or incomplete information. For example, the lack of appropriate language interpretation services can compromise a risk assessment.

Complete the B-SAFER in the following sequence of steps:

  1. Complete the background information. When possible record the case number, the names of the (alleged) offender and victim(s), and the sources of information consulted.

  2. Code the presence of individual risk factors. After all available information is reviewed code the presence of individual factors by filling in the appropriate symbol. A 4-point response format is used:

    • O = there is insufficient information available to code the factor;
    • N = the factor is definitely absent;
    • P = there is possible or partial evidence that the factor is present; and
    • Y = the factor is definitely present.

    Code each risk factor currently and in the past. "Currently" refers to the past 4 weeks, up to and including the incident under investigation.

  3. Rate risks for future spousal assault. Now consider the risk to intimate partners if NO INTERVENTION was taken. Make ratings for:

    • (a) imminent risk (less than two months);
    • (b) long-term risk (beyond 2 months); and
    • (c) risk for extremely serious assault or death.

    In each case risk is rated as Low, Moderate, or High (L, M, H).

  4. Devise a risk management plan. Finally, recommend actions based on the level of risk and specific risk factors that are present. The B-SAFER coding form organizes these actions into the following categories: monitoring, treatment, supervision, victim safety planning, and other considerations.

Definition of Risk Factors

The B-SAFER risk factors are described in detail on the following pages. We provide a brief rationale for each item's inclusion, as well as a definition to assist coding decisions. We have attempted to summarize key references, making note of relevant empirical reviews and professional guidelines.

1. Serious Physical/Sexual Violence

Rationale

Men who have demonstrated physically assaultive behaviour in either past or current intimate relationships are at risk for future intimate partner violence (Campbell, Sharps, & Glass, 2001; Dutton & Kropp, 2000; Fagan et al., 1983; Harrell & Smith, 1996; Healy, Smith, & O'Sullivan, 1998; Riggs, Caulfield, & Street, 2000; Saunders & Browne, 2002; Sonkin, 1987). Recidivism rate estimates for intimate partner violence range from 30 to 70 percent over a period of two years (Dutton, 1995); these rates seem to apply regardless of whether or not the offender is arrested or completes treatment (Gondolf, 2001; Hamberger & Hamberger, 1993).

In addition, typologies of spousal assaulters often indicate that the most severe patterns involve sexual assault (Gondolf, 1988; Snyder & Fruchtman, 1981). Men who have sexually assaulted their partners are also at greater risk of violent recidivism (Campbell et al., 2001; Goldsmith, 1990; Stuart & Campbell, 1989; Walker, 1989).

The significant recidivism of spousal assaulters may reflect patterns of behaviour learned in the assaulter's family of origin, as a significant number of these men experienced or witnessed violence as children (Caesar, 1988; Saunders, 1993; Schumacher, Feldbau-Kohn, Slep, & Heyman, 2000).

Coding

Y
The individual has physically/sexually assaulted a current or former intimate partner.
P
Possible or partial evidence that the individual has physically/sexually assaulted a current or former intimate partner.
N
The individual has not physically/sexually assaulted a current or former intimate partner.

Notes

"Physical/sexual assault" includes actual or attempted physical and sexual violence, including physical injuries and use of a weapon, but does not include threats (coded under item 2).

"Intimate partner" includes any wife, common-law spouse, or girlfriend.

2. Serious Violent Threats, Ideation, or Intent

Rationale

Thoughts or threats of causing harm to others are clearly relevant to risk assessment. It is common sense to consider threatening behavior when conducting a spousal violence risk assessment, but there is also empirical support for this risk factor. Men who make credible threats of death (i.e., men feared intensely by their partners) are at increased risk of violent recidivism (Gondolf, 1988; Sonkin, 1987; Dutton & Kropp, 2000; B. Hart, 1992; Stuart & Campbell, 1989; Walker, 1989). Also, spousal assaulters who have used or threatened to use a weapon are at increased risk for violent recidivism and spousal homicide (Campbell et al., 2001; Sonkin, Martin, & Walker, 1985).

Harassing or stalking behavior is a form of threatening that is of particular relevance to spousal violence. Stalking in the form of unwanted communicating, watching, following, or threatening can intentionally or recklessly create a sense of fear in victims. There is increasing evidence that such fear is justifiable given the link between the stalking of ex-intimate partners and violence (Burgess, et al, 1997; Kropp, Hart, & Lyon, 2002; Palarea, Zona, Lane & Langhinrichsen-Rohling, 1999; Douglas & Dutton, 2001). Stalking and threats are also risk factors for escalation into life-threatening violence (McFarlane, Campbell, & Watson, 2002). In general, any behavior or credible threat that generates significant fear in the victim should be considered relevant as some evidence suggests that such fear may be predictive of violence (Gondolf, 2001; Weisz, Tolman, & Saunders, 2000).

Overall, this factor is likely a risk marker that reflects the presence of mental illness, serious distress, or attitudes that support or condone intimate partner violence.

Coding

Y
The individual has serious violent ideation.
P
Possible or partial evidence that the individual has serious violent ideation.
N
The individual has no violent ideation.

Notes

"Violent ideation" includes thoughts, urges, and fantasies about killing or causing harm to others. It also includes intent, threats, or attempts to cause harm or death to others (including victim's friends or family members).

"Serious" means that the violent ideation is experienced as persistent and intrusive, involves high-lethality methods, or is associated with moderate to high intent.

Violent ideation may be inferred from behavior, as well as from threatening statements. Such inferences are more likely to be accurate when based on a pattern of behavior rather than a single act.

3. Escalation of Physical/Sexual Violence or Threats/Ideation/Intent

Rationale

Abusive relationships may be characterized by distinctive patterns or cycles of violence. One important pattern involves a recent escalation in the frequency or severity of assault. This pattern is associated with imminent risk for violent recidivism (B. Hart, 1992; Sonkin, 1987; Stuart & Campbell, 1989; Weisz, Tolman, & Saunders, 2000) and may reflect a "trajectory of violence"across time (Greenland, 1985). Escalation of intimate partner violence often is associated with life-threatening assaults (Campbell, 1995; Campbell et al., 2003).

Although it is not entirely clear why this pattern of violence occurs in some relationships but not others (Mahoney, Williams, & West, 2001), there may be a number of explanations for the escalation of violence in some relationships. For example, this pattern may reflect the instrumental, reinforcing aspects of the use of violence in intimate relationships. In other words, if the abuser obtains the outcome that he desires through violence, he will be more likely to use this strategy in the future. Escalation may also be related to desensitization to the use of violence over time, recent stressors, or the onset/recurrence of mental illness.

Coding

Y
The individual engages in physical/sexual violence or threats that escalate over time.
P
Possible or partial evidence that the individual engages in physical/sexual violence or threats that escalate over time.
N
The individual engages in physical/sexual violence or threats that do not escalate over time.

Notes

"Escalate" means the violence or threats have increased in severity or frequency over time. Increased severity indicates that, relative to earlier acts, the individual's recent acts of violence were more likely to involve direct contact with victims, serious physical harm to victims, or use of weapons or credible threats of death.

4. Violations of Civil or Criminal Court Orders

Rationale

There is abundant evidence in the literature that offenders who have violated the terms of conditional release (full parole, day parole, mandatory supervision, temporary absence) or community supervision (bail, probation) are more likely to recidivate than are other offenders (Andrews & Bonta, 1996, 2003; Hart, Kropp, & Hare, 1988; Nuffield, 1982). This relationship holds true when violent recidivism is the criterion (Quinsey, Harris, Rice, & Cormier, 1998).

Although there is little direct evidence bearing on this issue with respect to spousal violence specifically, based on the axiom that past behavior is a good predictor of future behavior, wife assaulters with a history of violating the "no contact" provisions of a civil or criminal court protective order (e.g., bail, probation, parole, restraining order, peace bonds) are likely to be at risk for violent recidivism.

This factor is a risk marker that may reflect generally antisocial attitudes, attitudes that support or condone intimate partner violence, severe distress, and employment or financial status. Thus, some research suggests that while protection orders are often helpful (Holt et al., 2003) abusers that have a lower stake in conformity are more likely to violate such orders (Carlson, Harris, & Holden, 1999; Sherman, Smith, Schimidt, & Rogan, 1992).

Coding

Y
Arrest(s) for violating the "no contact" provisions of a civil or criminal court order imposed because of spousal assault or to prevent spousal assault.
P
Violation(s) of the "no contact" provisions of a civil or criminal court order imposed because of spousal assault or to prevent spousal assault that did not result in arrest.
N
No violation of the "no contact" provisions of a civil or criminal court order imposed because of spousal assault or to prevent spousal assault, or the individual has never had such an order.

Notes

"Civil or criminal court order"includes bail, probation, parole, or restraining orders, as well as peace bonds and so forth.

5. Negative Attitudes About Spousal Assault

Rationale

It is often noted in the professional literature that most serious and persistent offenders minimize the seriousness of past violence, deflect personal responsibility for past violence, or even deny their involvement in past violence altogether. This is true of violent offenders in general and spousal assaulters in particular (Dutton, 1995; Dutton & Kropp, 2000; Hare, 1991; Riggs, Caulfield, & Street, 2000; Saunders, 1992; Webster et al., 1985).

In spousal assaulters, extreme minimization or denial is associated with an unwillingness to desist assaultive behaviour or to participate and complete treatment programs, which in turn is related to an increased risk of violent recidivism (Dutton, 1988, Gondolf & White, 2001; Hanson & Wallace-Capretta, 2000; Shepard et al., 2002; Sonkin, 1987). It is also plausible that minimization and denial will affect the degree to which an offender complies with other risk management strategies such as monitoring and supervision.

Research and clinical observation also suggest that a number of socio-political, religious, (sub-) cultural, and personal attitudes differentiate men who have recently assaulted their partners from those who have not (e.g., Andrews & Bonta, 1994; Campbell et al., 2001; Saunders, 1992b; Straus et al., 1980). For instance, spousal assaulters support or condone intimate partner violence by implicitly or explicitly encouraging patriarchy (male prerogative), possessiveness, misogyny, and/or the use of violence to resolve conflicts. These attitudes and beliefs are associated with increased risk of violent recidivism and femicide (Campbell et al., 2003; Daly & Wilson, 1998; Hanson & Wallace-Capretta, 2000; Sonkin, 1987; Riggs, Caulfield, & Street, 2000; Schumacher et al., 2001).

This factor may be causally related to future intimate partner violence given that attitudes have been shown to directly influence behaviour under certain circumstances (e.g., Ajzen & Fischbein, 1980). There is some evidence suggesting that these attitudes might be learned as a result of experiencing or witnessing family violence in childhood (Kessler, Molnar, Feurer, & Appelbaum, 2001; Riggs, et al., 2000; Schumacher et al., 2000).

Coding

Y
Explicitly endorses negative attitudes about spousal assault.
P
Appears to implicitly endorse negative attitudes about spousal assault.
N
No evidence of negative attitudes about spousal assault.

Notes

"Negative attitudes about spousal assault" include socio-political, religious, cultural or sub-cultural, and personal beliefs and values that directly or indirectly encourage or excuse abusive, controlling, and violent behavior. Such attitudes include sexual jealousy, misogyny, and patriarchy. Also included here is minimization or denial of violent actions or the serious consequences of those actions. Note that attitudes can be inferred from behavior (e.g., style of relating to women).