Abusive Behaviour Inventory (ABI): History
View Latest Version
Please note this is an old version of this entry, which may differ significantly from the current revision.
Contributor:
Rebecca Heron
,
Gracie McAndrew
,
Karen Parsonson
,
Kevin Browne

Abusive Behaviour Inventory (ABI) is regarded as an efficient self-report measure with demonstrated high reliability and validity.

  • ABI
  • domestic violence
  • intimate partner violence

1. Introduction

Domestic violence is a significant public health issue that spans the globe and relationship paradigms. Both mental and physical health repercussions are associated with domestic violence, including depression, suicide, injuries, and death (WHO 2013). The lifetime prevalence rate of intimate partner violence (IPV) data for women shows that 641 to 753 million women have experienced IPV since the age of 15 (WHO 2021). Researchers in the area of domestic abuse have highlighted the importance of assessing domestic violence to better assist public health (Graham et al. 2019). More specifically, researchers have stated that information needs to be obtained about the content types, the genders of perpetrator and victim, and methods to evaluate underreporting and overreporting (Follingstad and Rogers 2013).
The ABI is a domestic violence screening measure used to detect the presence of intimate partner violence. There are questions concerning physical abuse, psychological abuse, and sexual abuse. This tool aims to alert healthcare professionals to victims of domestic violence so they can provide safety resources. Shepard and Campbell (1992) are the original creators of the ABI. Within their paper, they include an examination of the ABI’s psychometric properties which provides data to suggest that the ABI is a respectable tool.

2. Criticisms of the ABI

The ABI is regarded as an efficient and effective domestic violence screening tool; however, it has been criticized for several reasons. Zink et al. (2007) criticized the tool for not providing clinical cut-off scores that would indicate a victim’s level of being physically or psychologically abused. Zink et al. (2007) investigated this and suggested a cut-off score of 10 for the two subscales (psychological/physical). Zink et al. (2007) rationalized that cut-off scores were important as they helped healthcare professionals and other professionals in their decision making when screening for domestic abuse. More recent literature has continued to hold the same opinion regarding the usefulness of assessments in conjunction with clinical judgement (Whiting and Fazel 2019).
Another criticism of the ABI is that it only focuses on abuse perpetrated or received within the last 6-month period. This may lead to an incomplete picture of violence the victim is suffering because some forms of violence are progressive or episodic (Ali et al. 2016; Ornstein and Rickne 2013). Short assessments such as the ABI are not likely to capture the actual extent of abuse as they only ask for abuse occurring within the last 6 months; therefore, any abuse prior to this will not be identified by the tool.
The ABI also fails to consider individual differences or circumstances for victims or abusers. For example, not all victims or abusers will have children. Yet, one of the questions on the ABI (item 11) assumes that all victims will have children, asking whether the perpetrator has tried to use their children against the victim. For victims who have no children, they will respond never, causing them to have a lower score for psychological abuse. However, this does not mean that the victim is not being psychologically abused in other ways. Other examples of the item’s lack of applicability to every victim include items that ask about work or school (item 22). Again, not all victims may go to work or school, and with the item (item 17) asking ‘has your partner driven recklessly when you were in the car’, it may be that their partner does not drive, but they may still be intimidating in other ways. Therefore, adaptations need to be made to the ABI’s items, as the currently existing items are likely to not apply to every victim and fail to take into account individual differences in circumstances.
Other criticisms of the ABI are that it is based on traditional feminist theory, which may limit the applicability of the measure. Traditional feminist theory builds on power, specifically patriarchal ideals of power which subordinate women. These ideals constitute the “Power and Control Wheel” (Cannon et al. 2015). The feminist theory of domestic violence is greatly responsible for IPV becoming a mainstream topic and inspiring policy changes. However, it is severely limited in its applicability and empirical support (Cannon et al. 2015). Thus, the ABI fails to provide a foundational understanding for LGBTQ and male victims of domestic violence. Walker et al. (2020) researched men’s experiences with IPV and their reporting behaviours. Their investigation revealed that up to 55.4% of men had experienced abuse from their female partners, and terms such as “boundary crossings” were used instead of domestic violence. The change in language served to reconcile the lack of identification men have with words such as “abuser” and “victim” that have gendered connotations. Men also noted barriers to reporting IPV, including fear of emasculation, police and care provider stereotyping, and not recognizing behaviours as abusive (Walker et al. 2020).
The ABI also fails to take into consideration situational motivations for abuse. Research indicates that abuse risk increases when partners become separated, and this violence can occur without previous abuse in the relationship (Ali et al. 2016). Further to this, violence upon separation is observed in both men and women and can increase with allegations of abuse (Ali et al. 2016). Therefore, the ABI is highly criticized for failing to consider the whole context, for example, the current state of the relationship and different motivations for violence.
The ABI is further criticized for not asking about both partners’ possible use of violence and abuse. Research has found that domestic violence can be mutual and in the form of resistance, such that violent resistance can increase abusive actions taken by the other partner (Ali et al. 2016). In addition, there is some evidence of the progression of IPV, and what may begin as predominantly male or female perpetration can evolve into mutual aggression (Leonard et al. 2014). Research concerning female perpetration seems to indicate that some forms may be retaliatory from being victimized, but this would only be one form of female-perpetuated abuse, as research has demonstrated that women can be both the instigators and equally aggressive as their partners (Ali et al. 2016).
For a very long time, the Conflict Tactics Scale (CTS) created by Murray Straus in 1979 was one of the most widely used scales to measure family violence (Straus 1979). However, this scale received extensive criticism, resulting in a second version being devised, the Conflict Tactics Scale Version 2. The CTS original version was criticized for many reasons, one of which was the combination of violent and non-violent acts of abuse within the same measure (Jones et al. 2017). The scale was further criticized for not focusing on the context of the situation, motives, and severity (Jones et al. 2017). Other criticism derived from the fact that this scale did not measure intimate partner violence specifically, but rather, was based on the conflict theory, which measured conflict in general. Therefore, it misrepresented domestic violence and failed to account for other elements such as controlling behaviours (Jones et al. 2017). The CTS-2, its second version, has received similar criticisms. The CTS-2 remains limited in its assessment of content types of IPV and situational contexts, still not accounting for perpetrator motivations (Jones et al. 2017). Such problems with the CTS and CTS-2 resulted in the development of the ABI. Shepard and Campbell (1992) devised this instrument to address the flaws that were identified in earlier questionnaires such as the CTS. Firstly, these measures differ in their theoretical underpinning. The CTS-2 has a conflict theory foundation, while the ABI is based on a coercive control model, including various abusive behaviours (Postmus et al. 2016). While the CTS-2 assumes a present conflict and acts of violence to be isolated incidents, the ABI includes behaviours that are intended to hurt and control partners (Postmus et al. 2016). This is a limitation of the CTS-2, as respondents are not given the opportunity to report abuse that is related to control or abuse not arising out of a known cause such as a conflict situation. For this reason, the ABI appears to be a better measure than the CTS and the CTS-2, as this tool does not assert that a conflict must occur for abuse to occur. Inclusion of more behaviours that take place within the context of domestic violence is an advantage because this increases the measure’s sensitivity, making it less likely to produce false negatives, and increases its applicability to more types of IPV.
Other questionnaires continue to be criticized due to their lack of assessment of multiple types of abuse and excluding psychological abuse. In a review of 10 IPV measures, only three tools assessed psychological abuse (Arkins et al. 2016). Those which screen for psychological abuse are the Abuse Assessment Screen (AAS), Humiliated Afraid Kicked and Raped (HARK), and Hurt Insulted Threatened or Screamed at Questionnaire (HITS), but HITS does not assess sexual assault (Arkins et al. 2016). Thus, Shepard and Campbell’s (1992) ABI instrument is supported not only for its ability to measure intimate partner violence, but also for specifically examining psychological abuse as this has not always been investigated by researchers. Another measure called the Composite Abuse Scale (CAS) covers items concerning physical, sexual, and psychological abuse (Ford-Gilboe et al. 2016). However, this measure aims to capture a woman-oriented experience of abuse (Ford-Gilboe et al. 2016). Therefore, it is not applicable to males and victims who do not identify as female. In addition, all of the measures mentioned are in need of additional research to further determine their validity and use with diverse populations (Arkins et al. 2016; Ford-Gilboe et al. 2016).

3. Psychometric Properties

According to the American Psychological Association’s (2017) Ethical Principles of Psychologists and Code of Conduct, the appropriate use of a psychometric assessment relies on its reliability, validity, and the population used to calibrate the test. Koocher and Keith-Spiegel, in their book Ethics in Psychology and the Mental Health Professions (Koocher and Keith-Spiegel 2016), assert that it is good practice for psychologists to examine psychometric properties of assessments to understand their strengths and limitations. The authors of the ABI did indeed examine the tool’s psychometric properties and provided data to show that the instrument does have good reliability and validity, and this is further discussed below.

3.1. Reliability

3.1.1. Internal Consistency

Shepard and Campbell (1992) reported the ABI to have good internal consistency following a study that they conducted in which they surveyed 100 men and 78 women. Males were selected from a chemical dependency treatment program; the females who participated were these males’ partners. They were each classified into one of four groups: females were either abused/not abused and males were either an abuser/not abuser. Relationship abuse was assessed during an interview with the males, and the females later completed an ABI questionnaire. Internal consistency was calculated using the alpha coefficient (Cronbach 1951). This manner of calculating internal consistency is the most common since the creation of the Cronbach alpha coefficient (Souza et al. 2017). Shepard and Campbell (1992) reported alpha coefficients for the four groups (abused/not abused; abuser/not abuser) ranging from 0.70 to 0.92. Therefore, the ABI was deemed to be very reliable, as the value for adequate to ideal internal consistency ranges from 0.6 to 0.7 or greater (Souza et al. 2017).

3.1.2. The Standard Error of Measurement

The standard error of measurement shows how much variability should be expected with repeated testing of the same person based simply on methods of measurement. Therefore, this is essential for the interpretation of scores from individuals (Polit 2015). Shepard and Campbell (1992) reported that the SEM scores for the four groups ranged from 0.04 to 0.12 in their study. Having a low standard error of measurement is crucial for domestic violence, as it informs risk assessment and risk management plans of individual patients.

3.1.3. Test Re-Test Reliability

The authors did not report test re-test reliability for the ABI. This is a weakness because the ABI only covers abuse within a 6-month period, and victims need to be continually assessed to monitor progress. Lacking information on this type of reliability is a gap in the literature.

3.2. Validity

3.2.1. Face Validity

Regarding the ABI, it can be concluded that this instrument does measure whether an individual has perpetrated or been a victim of physical/psychological abuse within a 6-month period. Therefore, it can be considered to have face validity. However, face validity is subjective and lacks scientific support; hence, the other areas of validity must also be considered.

3.2.2. Criterion Validity

Shepard and Campbell (1992) reported that the ABI has good criterion (concurrent) validity in their study. The established difference upon which they evaluated the criterion validity was the grouping of abused and non-abused participants. They found that mean scores for psychological and physical abuse in the abuser/abused group were 25% higher than in the non-abuse groups. For males, the difference in group means was 0.55 (psychological abuse subscale) and 0.42 (physical abuse subscale) and the differences for the females’ group means was 0.80 (psychological abuse subscale) and 0.55 (physical abuse subscale). Therefore, the ABI did show good criterion validity in both cases and significant differences were observed at the 0.001 level.

3.2.3. Construct Validity

In their study, Shepard and Campbell (1992) reported the ABI to have good construct validity, possessing both convergent and discriminant validity. They found variables that are highly related to abuse in relationships (such as clinical assessment of abuse, client assessment of abuse, and previous arrest for domestic violence) to be significantly correlated with the subscales for both men and women. In addition to this, they found variables believed not to correlate as highly with domestic abuse; for example, household status and age did not correlate as highly with the subscales for men and women as did the other variables.
Additional research conducted by Zink et al. (2007) compared the ABI to the CTS-2. The correlations of the ABI and CTS-2 total scores were 0.76 with significant differences at the 0.001 level (Zink et al. 2007). The correlations of the physical abuse subscale of the ABI with the verbal physical aggression, injury, and sexual coercion subscales of the CTS-2 were 0.71 with significant differences at the 0.001 level (Zink et al. 2007). Similarly, correlations between the psychological abuse subscale of the ABI with the verbal aggression scale of the CTS-2 were 0.74 with significant differences at the 0.001 level (Zink et al. 2007). This illustrates that the ABI has adequate criterion validity and is comparable to a well-established measure such as the CTS-2.

3.2.4. Factor Validity

Shepard and Campbell (1992) examined the factorial validity of the ABI by computing the correlations between each item. They found that although many items were highly correlated with related variables, some were correlated with unrelated variables, behaviour, or being violent towards objects. Therefore, it made sense that these items correlated with both the psychological and physical subscale. Other modifications to the test involved the omission of item 21 (being spanked), as a result of this item having negative connotations and receiving a low response rate. For example, 93% of participants in their study stated that they had never been spanked or spanked someone else. Due to these modifications, it may be argued that the ABI does have factorial validity as the authors examined the factorial validity of the test and updated the test accordingly.

3.3. Appropriate Norms/Populations

In order to gain an accurate interpretation of a psychometric measure, appropriate norms are required. Shepard and Campbell (1992) provide norms for the ABI; however, these norms are based on a US based sample and therefore they may not be generalizable to cultures outside the US. As it stands, there is no evidence to suggest that the ABI has been culturally validated and so this is problematic when using the tool with individuals from non-Western countries. Domestic violence researchers have stressed the importance of multicultural considerations in progressing the area, as culture can influence risk assessment and safety predictions (Mallory et al. 2016).
Further to this, Shepard and Campbell (1992) also based the norms on an in-patient chemical dependency program, hence again, the norms are limited as they may not apply to normal populations. Research shows that substance and alcohol use have an influence on both the perpetration and victimization of IPV (Cafferky et al. 2018). Additionally, substance and alcohol use have different effects depending on gender, such that male users are more likely to perpetrate violence and female users are more likely to be victimized (Cafferky et al. 2018). Therefore, as the relationship is complex between drug and alcohol use, IPV perpetration, victimization, and gender, the ABI should be standardized on a sample that is not in a chemical dependency program.
In the study conducted by Zink et al. (2007), they analysed the false negative rate of the ABI. They discovered that those who were not identified as a victim of abuse were younger African-American women. This is crucial to note as mental health providers must be aware of a measure’s tendency to under report in certain populations.

3.4. Distorted Responses

Self-reporting bias, social desirability bias, and recall bias have been reported to be confounding factors in self-report tools (Althubaiti 2016). Although the ABI relies on self-report, a validity scale was not included by the authors. However, the authors did advise that male respondents’ (potential perpetrators’) scores should be interpreted with caution as they may respond defensively. Research illustrates that perpetrators of domestic violence often minimize and deny their abusive behaviours, more specifically underreporting, further highlighting the importance of accurately assessing domestic violence potential’s role to better assist public health (Graham et al. 2019). In Strang and Peterson’s (2020) study, when men were led to believe the honesty of their responses would be evaluated, they admitted up to 6.5 times more sexual assault behaviours. This suggests that the presence or perceived presence of validity scales can produce more candour in respondents. Conversely, the ABI tool may be advantageous in that it may be able to assess for defensive responses if both the victim and the perpetrator complete the tool. This may allow for their results to be compared and any inconsistencies in the victim’s/perpetrator’s responding may be explored. This could also be used in the context of treatment, allowing clinicians to see if there is a lack of insight/denial of abuse on the part of the victim/perpetrator.
Furthermore, research has shown that female victims are not always able to recognize their partners’ subtly controlling behaviours as abuse, and medical professionals lack awareness on the subject as well as training on how to approach the issue (Bradbury-Jones et al. 2014). This suggests that female victims’ responses may not be fully accurate as they may respond with low scores on the psychological abuse subscale because of their inability to recognize this as abuse. Therefore, it is advised that when using the ABI, if professionals perceive that an individual is responding in a biased manner, they should perhaps use additional tests to assess for distorted responding. Not doing so could expose the victim to further abuse (Bradbury-Jones et al. 2014), and according to the American Psychological Association’s (2017) Ethical Principles of Psychologists and Code of Conduct, a mental health professional should always consider these factors during interpretation of an assessment.

3.5. Structural Analysis

The creators of the ABI-R conducted a confirmatory factor analysis over the items of the ABI to determine if the two-subscale model was sufficient. They found that the conceptualization of IPV into physical and psychological abuse proved to be an inadequate structure for the ABI (Postmus et al. 2016). In conducting their CFA, they used a cut-off for a satisfactory model fit of CFI = 0.90 (Postmus et al. 2016; Byrne 2001). However, the statistical analysis revealed a value of CFI = 0.775., indicating that a two-factor model was not a bad fit for the measure. These results supported the creation of a third subscale of sexual abuse in the ABI-R (Postmus et al. 2016).

This entry is adapted from the peer-reviewed paper 10.3390/socsci12070400

References

  1. WHO (World Health Organization). 2013. Global and Regional Estimates of Violence against Women: Prevalence and Health Effects of Intimate Partner Violence and Non Partner Sexual Violence. Geneva: World Health Organization.
  2. WHO (World Health Organization). 2021. Violence against Women Prevalence Estimates, 2018: Global, Regional and National Prevalence Estimates for Intimate Partner Violence against Women and Global and Regional Prevalence Estimates for Non-Partner Sexual Violence against Women. Geneva: World Health Organization.
  3. Graham, Laurie, Kashika Sahay, Cynthia Rizo, Jill Messing, and Rebecca Macy. 2019. The validity and reliability of available intimate partner Homicide and Reassault Risk Assessment Tools: A Systematic Review. Trauma, Violence, and Abuse 22: 18–40.
  4. Follingstad, Diane, and Jill Rogers. 2013. Validity concerns in the measurement of women’s and men’s report of intimate partner violence. Sex Roles: A Journal of Research 69: 149–67.
  5. Shepard, Melanie, and James Campbell. 1992. Abusive behavior inventory. Journal of Interpersonal Violence 7: 291–305.
  6. Zink, Therese, Lisa Klesges, Linda Levin, and Frank Putnam. 2007. Abuse behavior inventory: Cutpoint, validity, and characterization of discrepancies. Journal of Interpersonal Violence 22: 921–31.
  7. Whiting, Daniel, and Seena Fazel. 2019. How accurate are suicide risk prediction models? Asking the right questions for clinical practice. Evidence Based Mental Health 22: 125–28.
  8. Ali, Parveen, Katie Dhingra, and Julie McGarry. 2016. A literature review of intimate partner violence and its classifications. Aggression and Violent Behavior 31: 16–25.
  9. Ornstein, Petra, and Johanna Rickne. 2013. When does intimate partner violence continue after separation? Violence Against Women 19: 617–33.
  10. Cannon, Clare, Katie Lauve-Moon, and Fred Buttell. 2015. Re-theorizing intimate partner violence through post-structural feminism, queer theory, and the sociology of gender. Social Sciences 4: 668–87.
  11. Walker, Arlene, Kimina Lyall, Dilkie Silva, Georgia Craigie, Richelle Mayshak, Beth Costa, Shannon Hyder, and Ashley Bentley. 2020. Male victims of female-perpetrated intimate partner violence, help-seeking, and reporting behaviors: A qualitative study. Psychology of Men & Masculinities 21: 213–23.
  12. Leonard, Kenneth, Jamie Winters, Jill Kearns-Bodkin, Gregory Homish, and Audrey Kubiak. 2014. Dyadic patterns of intimate partner violence in early marriage. Psychology of Violence 4: 384–98.
  13. Straus, Murray. 1979. Measuring Intrafamily Conflict and Violence: The Conflict Tactics (CT) Scales. Journal of Marriage and Family 41: 75–88.
  14. Jones, Richard, Kevin Browne, and Shihning Chou. 2017. A critique of the revised conflict tactics scales-2 (CTS-2). Aggression and Violent Behavior 37: 83–90.
  15. Postmus, Judy, Amanda Stylianou, and Sarah McMahon. 2016. The abusive behavior inventory–revised. Journal of Interpersonal Violence 31: 2867–88.
  16. Arkins, Brigid, Cecily Begley, and Agnes Higgins. 2016. Measures for screening for intimate partner violence: A systematic review. Journal of Psychiatric and Mental Health Nursing 23: 217–35.
  17. Ford-Gilboe, Marilyn, Nadine Wathen, Colleen Varcoe, Harriet MacMillan, Kelly Scott-Storey, Tara Mantler, Kelsey Hegarty, and Nancy Perrin. 2016. Development of a brief measure of intimate partner violence experiences: The composite abuse scale (revised)—Short form (CAS r -SF). BMJ Open 6: 1–13.
  18. American Psychological Association. 2017. Ethical Principles of Psychologists and Code of Conduct. Amended Effective 1 June 2010, and 1 January 2017. Available online: https://www.apa.org/ethics/code/ (accessed on 14 March 2023).
  19. Koocher, Gerald, and Patricia Keith-Spiegel. 2016. Ethics in Psychology and the Mental Health Professions: Standards and Cases. Oxford: Oxford University Press.
  20. Cronbach, Lee. 1951. Coefficient alpha and the internal structure of tests. Psychometrika 16: 297–334.
  21. Souza, Ana, Neusa Alexandre, and Edineis de Guirardello. 2017. Psychometric properties in instruments evaluation of reliability and validity. Applications of Epidemiology 26: 649–59.
  22. Polit, Denise. 2015. Assessing measurement in Health: Beyond reliability and validity. International Journal of Nursing Studies 52: 1746–53.
  23. Mallory, Allen, Prerana Dharnidharka, Sharon Deitz, Patricia Barros-Gomes, Bryan Cafferky, Sandra Stith, and Kimberly Van. 2016. A meta-analysis of cross-cultural risk markers for intimate partner violence. Aggression and Violent Behavior 31: 116–26.
  24. Cafferky, Bryan, Marcos Mendez, Jared Anderson, and Sandra Stith. 2018. Substance use and intimate partner violence: A meta-analytic review. Psychology of Violence 8: 110–31.
  25. Althubaiti, Alaa. 2016. Information bias in health research: Definition, pitfalls, and adjustment methods. Journal of Multidisciplinary Healthcare 9: 211–17.
  26. Strang, Emily, and Zoe Peterson. 2020. Use of a bogus pipeline to detect men’s underreporting of sexually aggressive behavior. Journal of Interpersonal Violence 35: 208–32.
  27. Bradbury-Jones, Caroline, Julie Taylor, Thilo Kroll, and Fiona Duncan. 2014. Domestic abuse awareness and recognition among primary healthcare professionals and abused women: A qualitative investigation. Journal of Clinical Nursing 23: 3057–68.
  28. Byrne, Barbara. 2001. Structural Equation Modeling With AMOS, EQS, and LISREL: Comparative Approaches to Testing for the Factorial Validity of a Measuring Instrument. International Journal of Testing 1: 55–86.
More
© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.
This entry is offline, you can click here to edit this entry!
Video Production Service
Feedback