By Frank R. Ascione, Ph.D, Claudia V. Weber, M.S., and David S. Wood,
Utah State University, Logan, Utah
Originally submitted to The Geraldine R. Dodge Foundation, April 25, 1997
"I have spoken with the most vulnerable and seen how poisoned they have become... They show us that committing acts of violence is traumatic. They show us that witnessing acts of violence is traumatic. And they show us that inside most of the adolescent and adult perpetrators of violence are traumatized children, untreated children, frightened children."OVERVIEW
—James Garbarino (1995)
Raising children in a socially toxic environment. San Francisco: Jossey-Bass. (p.86)
This study represents the first attempt to assess animal maltreatment in samples of women seeking safety at shelters for women who are battered (101 women) and community samples of women who were not battered (60 women). We developed instruments to measure pet abuse from women and children's reports and assessed levels of family violence (not including pet abuse) using the well known Conflict Tactics Scale. Current and past pet ownership (in the past 12 months) was high for all samples (ranging from 64.1% to 96.6%) but the shelter sample reported lower current pet ownership than the community sample. There were also lower levels of regular and emergency veterinary care and pet vaccination rates in the shelter sample. Pet "turnover" over the past 5 years was higher in the shelter group. Partners (e.g., husbands, boyfriends) of shelter women were less likely to help care for animals than partners of community sample women.
Shelter women were more likely to report that their partners had threatened to hurt their pets (52%) than community sample women (16.7%). The severity of these threats was also higher in the shelter sample. Actual hurting or killing of pets was reported by 54% of the shelter women but only 3.5% of the community women. In the majority of cases, shelter women reported that multiple incidents of hurting or killing pets had occurred. In the shelter group, nearly one in four women reported that concern for their pets had kept them from coming in to the shelter sooner.
Approximately half of the shelter women reported that their children had witnessed pet abuse in contrast to less than four percent of the community sample women. One in four shelter group women and one in five community group women reported that one of their children had hurt or killed pets. The severity of hurting pets was lower, however, for the community group children. Shelter group children also displayed higher behavior problem scores on a standardized measure than community group children.
Shelter group children (N=39) were also interviewed directly and two-thirds reported witnessing pet abuse (including the strangling, poisoning, or shooting of a pet). Nearly half (46.4%) of the incidents involved the father, stepfather, or woman's boyfriend as the perpetrator. The severity of the abuse was rated as involving pain or discomfort to the animal or rated as torturing or killing of the animal by 88.5% of the children reporting witnessing of pet abuse, with 92.6% of the children indicating they were "sort of upset " or "very upset" by the incidents. Although 92.1% of the children said they helped take care of pets, l 3.2% admitted to hurting pets and 7.9% reported hurting or killing other animals. Nevertheless, 51.4% of these children said they had protected pets, in some cases by directly intervening to keep pets from being harmed.
Over two-thirds of the children (67.6%) said they would like to see pets in their home treated "better than they are now treated".
Use of the Conflict Tactics Scale verified that the levels of minor and severe physical aggression were substantially higher in the shelter than in the community groups. The highest levels of severe physical aggression (e.g., beating, burning, threatening with a gun or knife) occurred in the shelter groups when either threats to or hurting/killing of pets was present in these households.
The results of this study illustrate the landscape of terror in which many women, children, and their pets reside and should prompt renewed attention to the confluence of family violence and animal maltreatment.
PURPOSE OF THE PROJECT
As I noted in the original proposal, increasing attention to the issue of animal maltreatment has led to examination of the contexts in which such maltreatment takes place. The limited research that does exist has concentrated on the associations between animal abuse and either child physical abuse or sexual abuse. Pet abuse in families where domestic violence (adult partner battering) is present has received significantly less research scrutiny.
With many animal welfare organizations beginning to collaborate with human welfare programs to address violence (its origins, prevention, and reduction), data are needed to document the coexistence of interpersonal aggression and the maltreatment of animals, especially pets which enter into significant relationships with their human caretakers. Such data would highlight the scope of the problem and suggest areas for effective cooperation between animal and human welfare agencies (e.g., boarding pets when a woman must flee her home for protection). These data could also be used to facilitate legislative changes regarding the seriousness of violence toward animals. Finally, children's exposure to the maltreatment or killing of their beloved pets can be considered a form of terrorizing or psychological abuse. It is of interest to note that the Utah legislature recently passed a bill making domestic violence offenses more serious when a child has witnessed partner battering. Witnessing a pet being abused or tortured may be no less distressing to children.
Only one published study (Ascione, 1997) has specifically addressed the presence of animal maltreatment in a sample of women who were battered. The sample was small (38 women) and the study lacked a comparison group of women who were not battered by their partners. In addition, only the women were interviewed so children's perspectives on pet abuse were not examined. This study also failed to measure the frequency and intensity of domestic violence. The current research was designed to correct these research flaws and limitations on generalizations from the earlier results.
We collaborated with five Utah crisis shelters for women who are battered to enlist participation by women who had experienced domestic violence. We also sought participation by women, in the community, who did not have a history of domestic violence. The main criteria were that women currently have pets or had pets in the past year and, for part of the sample, that they had a child between 5 and 18 years of age who would be willing to participate. The Appendix contains the protocol used for selecting participants as well as samples of the informed consent documents used with participants. These materials as well as the overall project protocol had been approved by Utah State University's Institutional Review Board for research with human participants and the state agency overseeing shelter operations. It should be emphasized that all participants were volunteers.
Shelter participants were invited to take part in the project by shelter staff and community participants were enlisted through newspaper ads and posters placed throughout the community. Our initial plan was to secure the shelter sample and then match its demographic characteristics with the community sample. However, the prolongation of securing the shelter participants (due to director and staff turnover at shelters) required us to begin testing community volunteers before the majority of shelter participants were tested. The generous time extensions granted by the Dodge Foundation did allow us to meet the target sample size and, with assistance from Utah State University's Vice President for Research, we were able to increase the community sample's size beyond that specified in the original proposal.
To facilitate reporting sample characteristics and results, I will use the following abbreviations throughout the remainder of this report. Participants were categorized into these four groups:
- S-C This group included women entering shelters for women who are battered and who agreed to allow interviewing of one of their children. The number (N) who were assessed was 39.
S-NC This group included sheltered women who did not have children, had children whose ages did not meet our selection criteria, or who elected to not include their children in the project. N=62.
NS-C This group included community women without a history of domestic violence who agreed to report on one of their children (community children were not interviewed due to the potentially disturbing nature of some of the assessment questions). N=30.
NS-NC This group included community women without a history of domestic violence. None of these women had children. N=30.
The age range of participants was 17 to 57 years with the following breakout for groups' mean age: S-C 34.05 yrs; S-NC 30.23 yrs; NS-C 40.33 yrs; NS-NC 25.67 yrs. Although the mean age of the S-C group was lower than the NS-C group and higher for the S-NC than NS-NC group, the overall mean age of the shelter women (31.7 yrs) was comparable to that of the community women (33 yrs.).
Reported marital status for the entire sample was: married - 64%, divorced - 9.9%, single - 24.8%, and widowed -1.2%. There was greater diversity of marital status in the shelter samples with more women falling into the divorced, single, and widowed categories.
Ethnic status was also more varied in the shelter samples but closely matched demographics from a state-wide study of women who had been battered (Thompson, 1994). For the S-C group, 71.8% were European American, 17.9% Hispanic, 2.6% Native American, 5.1 % African American, and 2.6% "other" (e.g., mixed, Samoan). The corresponding figures for the S-NC group were 66.1%, 9.7%, 9.7%, 9.7%, and 4.8%. The community samples were overwhelmingly European American but the NS-C group included 6.7% Native Americans.
The number of years of education women reported for themselves and their partners was lower, by two or three years, for shelter than community group women. The mean years of partners' education for the S-C, S-NC, NS-C, and NS-NC groups were 11.83, 11.80, 15.5, and 14.4, respectively. Corresponding figures for women were 12.61, 12, 14.57, and 15, respectively.
We used information about education and type of employment in computing Hollingshead ratings of social status. In general, the shelter group scored lower than the community group on this index. However, the results varied depending on whether both a woman and her partner were employed, only the woman was employed, or only the partner was employed. The group scores fell primarily in the semiskilled worker to minor professional / technical worker categories. Specific Hollingshead scores are listed in the Appendix and will be reported when this research is prepared for submission to an academic journal.
Shelter women and children were interviewed by shelter staff once the initial crisis of entering the shelter had subsided. Community women were interviewed by a psychology graduate assistant. Women were informed about the nature of the project, the type of assessments, methods for insuring the confidentiality of data, and the payment they would receive. Informed consent forms and payment receipts were kept at the shelters to maintain the confidentiality of shelter participants. Community participant forms and receipts were kept in locked university files.
The primary assessments were the 30-item Battered Partner Shelter Survey (BPSS)/Pet Maltreatment Survey (Ascione & Weber, 1996 revised) and a similar Families and Pets Survey (PAPS; Ascione & Weber, 1996) for use with community women, the 17-item Children's Observation and Experience with Their Pets (COEP) Survey (Ascione & Weber, 1995) used only with the S-C group, and the 19-item Conflict Tactics Scale (CTS; Straus, 1993) used with all adult participants. Copies of these instruments are included in the Appendix. In addition, S-C and NS-C women completed the 113-item Child Behavior Checklist (Achenbach, 1991) on "the child who has the most contact (positive or negative) with pets" in their families. The content of these instruments will be elaborated in the Results section of this report.
Completed assessments were transformed into computer-ready data using a 24-page codebook. In addition, open-ended comments and other qualitative information were transcribed and used for descriptive purposes (e.g., other pet-related issues women chose to share with the interviewers).
Information about demographics, history of pet ownership, pet care, and violence toward pets was assessed with the BPSS and the PAPS.
We found that virtually all the women had pets in the past 12 months (SC 92.1%; S-NC 90.3%; NS-C 96.6%; NS-NC 100%), but shelter women were less likely to currently own a pet (S-C 64.1%; S-NC 70.5%; NS-C 90%; NS-NC 83.3%). When women were asked about the number of pets they had over the past five years, the mean number of pets was generally higher for the two shelter groups (S-C 9.19 pets; S-NC 6.15; NS-C 5.6; NS-NC 5.4) suggesting a higher rate of pet "turnover" in the shelter groups. If one deletes the highest outlier figure for number of pets owned in a five-year period, the ranges for the four groups differ remarkably: S-C 1-40 pets; S-NC 1-30; NS-C 1-16; NS-NC 1-13.
Reports of veterinary care for pets also differed for the shelter and community groups. Pets were reported to receive regular veterinary care by 55.4% of the shelter women, but by 85% of the community women. Emergency veterinary care was reported by 31.7% of shelter and 48.3% of community women. Pets were reported to have received their vaccinations by 73% of shelter and 93.1 % of community women. These data suggest generally lower levels of veterinary care for pets in families where there is domestic violence.
When asked whether their partners (e.g., husbands, boyfriends) helped to care for pets in their homes, shelter women were less likely to respond positively. The percentage "Yes" responses to this question for the four groups were: S-C 51.3%; S-NC 69.4%; NS-C 86.7%; NS-NC 86.7%.
A number of items asked about the variety of care for pets provided by partners. When asked if the only care their partner provided was to feed pets, "Yes" responses were as follows: S-C 50%; S-NC 46.7%; NS-C 11.5%; NS-NC 3.8%. Partners of community women were clearly more involved in a variety of pet-related activities beyond feeding, including walking, grooming, and playing with pets. Partners' caring for pets by taking them to the veterinarian was also more common in the community samples: S-C 0%; S-NC 4.4%; NS-C 15.4%; NS-NC 19.2%.
Threats Toward Pets
When women were asked whether their partners had ever threatened to hurt their pets, more shelter women (52%) than community women (16.7%) responded affirmatively. The percent responding "Yes" for each group was: S-C 52.6%; S-NC 51.7%; NS-C 20%; NS-NC 13.3%. Although there was a difference between groups in the presence of threats, differences in the severity of threats (the majority were threats to seriously abuse or kill a pet) and their frequency was comparable across groups (e.g., 50 to 55% of women in each group who had reported threats indicated that threats were repeated "multiple times").
Women who had mentioned threats were asked to report their judgments about what the threats were related to. Shelter women were less likely to say that threats were related to the pets' actions or behaviors (S-C 15.8% and S-NC 25%) than community women (NS-C 83.3% and NS-NC 50%). None of the community women indicated that threats were related to their partners' attempts at coercion; however, 21.1% of S-C and 15.6% of S-NC women reported that this was the case.
Actual Harm to or Killing of Pets
Here again, there was a substantial difference between the reports of shelter and community women. When asked if their partners had ever actually hurt or killed one of their pets, 69.2% of S-C, 44.3% of S-NC, 7.1% of NS-C, and 0% of NS-NC group women said "Yes." Regardless of group membership, some women indicated that pets had been hit or kicked, or had been shot. The more horrific instances seemed to be restricted to the reports of shelter women who reported the following examples (among many others): pet was drowned, pet was nailed to the woman's bedroom door, pet was given alcohol and poison, pet's entire fur coat was shaved during the winter, and pet was thrown out of a moving car. Most of the incidents involved cats or dogs, but in the shelter groups, birds, gerbils, and rabbits were also mentioned as victims of abuse or killing. Although community women indicated that actual harm or killing only occurred once, 70% of S-C and 52.6% of S-NC group women indicated such incidents occurred "multiple times." Of the women reporting, 83.3% of S-C and 88.9% of S-NC groups noted that they were "very close" to the pet that was harmed.
Although community women cited the pet's biting or "discipline" as the reason pets were harmed, most of the shelter women (70.4% for S-C and 60.7% for S-NC) could not give a reason why pet abuse occurred. Those in the shelter groups who did provide a reason often mentioned either the animal's behavior, the partner's attempt at coercion, or both.
Four of the shelter group women and two community group women indicated that they themselves had hurt a pet. Except for one NS-NC group woman who said she once killed an animal ("At a time in my life when I was sick"), the other women described the incidents as accidents, mainly involving vehicles.
Women's Ratings of Their Emotional Responses
When asked to rate how they felt after their pets were threatened, shelter women were more likely to respond that they felt "numb" or "terrible" (90% for S-C and 87.6% for S-NC) than community women (33.3% for NS-C and 75% for NS-NC). Similar ratings were provided when women were asked about their feelings when a pet was actually hurt or killed (percent responding that they were "numb" or felt "terrible"—S-C 85.7%; S-NC 92.8%; NS-C 50%; NS-NC no instances of abuse/killing were reported).
Although none of the community women said that they were relieved that their pet had been threatened instead of themselves, 5% of S-C and 15.6% of S-NC women said they were relieved. Similarly, none of the community (NS-C) women said they were relieved that their pet had been hurt instead of themselves, but 10.7% of S-C and 3.4% of S-NC women said they were relieved.
Concern For Pet Welfare
Only shelter women were asked the following question: "Did concern for your pet keep you from coming in to the shelter sooner?" Nearly one in four shelter women (23.1% for S-C and 22.6% for S-NC) responded "Yes" to this item. We also examined responses to this question in relation to whether or not women had reported threats or harm to pets and found even higher percentages. For the S-C group, if pets had been threatened or if they had been hurt, 35% and 25.9%, respectively, of the women said they had delayed entering the shelter. The figures for S-NC women were 25.8% and 29.6%.
Women were also asked whether their partners' willingness to use violence toward the women, their children, and their pets had changed during the time they were with their partners. For S-C women, 59% and for S-NC women, 69.4% indicated that their partners' violence toward them and their children had increased. Increases in willingness to use violence toward pets was reported by 30.8% of S-C and 29% of S-NC women.
The next section describes results from women's reports on their children's behavior with pets. Recall that S-C and NS-C groups were asked to select a child who had the most contact (positive or negative) with pets (herein referred to as "target children"). Shelter women (both SC and S-NC) were also asked to report on children who were not targeted for additional assessment in this research or, for the S-NC group, children women did not want included in the study. Also, recall that none of the women in the NS-NC group had any children.
Children's Caring For Pets
The overwhelming majority of target children were reported to take care of pets (94.9% for S-C and 96.7% for NS-C). Caring for pets was somewhat lower for non-target children in the shelter groups (87.9% for S-C and 78.9% for S-NC). These data and data to be reported below suggest that S-C women may have selected target children who had better relations with pets than other children in the family.
Children's Observation of Pet Abuse
When asked whether their target children had ever observed pet abuse in their homes, 61.5% of S-C mothers in contrast with 3.3% of NS-C mothers responded "Yes." For non-target children, 51.5% of S-C and 34.2% of S-NC were reported to have witnessed pet abuse.
Children's Abuse of Pets
When asked whether target children had ever hurt a family pet, 10.5% of S-C and 20% of NS-C women said that they had. Most of the NS-C mothers (60%) rated the severity of the harm done as "minor teasing;" most of the S-C mothers (66.7%) rated the severity as "caused painful discomfort" or "killed." Examples included restraining a pet, breaking a pet's leg, and suffocating a pet.
When asked about non-target children's abuse of pets, 27.3% of S-C and 34.2% of S-NC mothers indicated that these other children had hurt pets. These data, along with data cited earlier, suggest that S-C mothers may have selected target children who had fewer problems in their relations with family pets.
Children's Child Behavior Checklist (CBCL) Scores
Maternal reports on children's behavior problems using the CBCL indicated higher Total Behavior Problems, Internalizing (e.g., anxiety-related items), and Externalizing (e.g., aggressive, acting out items) Scores for the S-C than the NS-C group (mean scores for these three scales were 61.9 vs 52.3, 62.1 vs 52.4, and 59.7 vs 51.4 for the S-C and NS-C groups, respectively).
We also examined the percentage of children in the S-C and NS-C groups whose CBCL scores fell into the clinically significant range for behavior problems. The data for the Total Behavior Problems, Internalizing, and Externalizing Scores, respectively, were: S-C 38.5%, 48.7%, and 38.5%; NS-C 13.3%, 13.3%, and 10%. Thus, more shelter children's CBCL scores fell into the clinical range on this assessment.
The next section of the results reports on S-C group target children's own responses to the interview using the COEP. Thirty-nine children were interviewed whose ages ranged from 5 to 18 years (the mean age was 9.9 years). Forty-three point six percent were girls and the children reported either no siblings or up to seven siblings. Current pet ownership was acknowledged by 52.6% of the children and pet ownership in the past 12 months by 92.3% of children, which corresponds well with maternal reports.
Observation of Pets Being Threatened or Hurt
Children were asked if anyone had ever said they would hurt or kill one of their pets, and 40% of the children acknowledged hearing such threats. The severity of the threats included ratings of serious abuse, torture, or killing for 78.5% of these children. When asked whether they had ever seen or heard their pets being hurt, 66.7% of the children said "Yes." they described witnessing poisoning, strangling, leaving a pet out in the cold, and the shooting of a pet.
With regard to the severity of the observed pet abuse, 11.5% of the children described it as "annoying or frightening" the pet, but 88.5% said the abuse entailed pain, discomfort, torturing, or killing the pet. Although 28.6% of the children could not identify who had hurt their pets, 46.4% identified either their fathers, their stepfathers, or their mothers' boyfriends as the perpetrator of the abuse. When asked, "How did you feel after the pet was hurt?" 92.6% of the children said they were either sort of upset (33.3%) or very upset (59.3%).
Children Hurting Pets
When asked if they had ever hurt a pet, 13.2% of the children said that they had and then described incidents of throwing, hitting, or stepping on a pet. Of these five children, three indicated that the severity of harm involved annoying or frightening a pet; the other two children said that the pet had been killed. When asked about animals other than pets that they had hurt or killed, 7.9% of the children admitted to such acts.
Children Caring For Pets
Children were asked if they had ever taken care of a pet, and 92.1 % said that they had (which corresponds to maternal reports). All of the children said that they had a favorite pet with 79% mentioning dogs and/or cats; birds, rabbits, guinea pigs, and reptiles were also mentioned as favorites by the remaining children.
We also asked children if they had ever protected a pet and 51.4% said they had. The interventions that children described were varied and included keeping the pet in their room (5.9%), moving the pet to a different location, presumably in their home (47.1%), and blocking the animal from a perpetrator's assault (29.4%).
The final question asked of children was how they would like to see pets treated in their homes. Over two-thirds (67.6%) of the children said they would like to see pets treated "better than they are now treated."
The final component of this results section describes data we derived from the Conflict Tactics Scale (CTS). The CTS is a widely used measure of family violence and we incorporated it in this study to verify that the shelter groups' experience of domestic violence was different from the community groups'. We also wanted to examine the relation between CTS scores and pet abuse.
Measure of Family Violence
The CTS is a 8-item rating scale introduced with the following question: "When you and your partner have a problem, what sort of things have you done to solve it?" All women in the study completed the CTS twice, first rating their own tactics and then rating their partners' tactics.
The items range from VERBAL TECHNIQUES (e.g., "discussed an issue calmly"), VERBAL AGGRESSION (e.g., insults, swearing), MINOR PHYSICAL AGGRESSION (e.g., pushing, grabbing, shoving), to SEVERE PHYSICAL AGGRESSION (e.g., burned, scalded, used a knife, fired a gun). These four subscales yield different levels of mean scores since the number of items in each subscale differ and some subscales, for example Severe Physical Aggression, use a weighting procedure. Each item is scored as to the tactic's presence and its frequency of use. Except for the Verbal Techniques subscale, higher scores reflect higher levels of problematic verbal or physical aggression.
The following table lists the mean CTS subscale scores for women's self-ratings and ratings on their partners for each of the four groups.
|Women's reports on self||31.4||26.6||23.6||27.9|
|Women's reports on partners||12.2||12.2||23.1||21.8|
|Women's reports on self||78.7||78.9||21.8||23.6|
|Women's reports on partners||102.1||100.1||15.8||18.5|
|Minor Physical Aggression|
|Women's reports on self||10.6||11.1||.5||.5|
|Women's reports on partners||35.6||34.1||.5||1.3|
|Severe Physical Aggression|
|Women's reports on self||6.6||8.1||.13||.1|
|Women's reports on partners||148.7||142.7||.2||.83|
These data suggest that shelter women and community women were similar in their use of Verbal Techniques during problem solving, but that partners of shelter women used these techniques at a lower rate. The remainder of the table verifies that the shelter groups scored higher on Verbal Aggression, Minor Physical Aggression, and Severe Physical Aggression than the community groups. This was true for both the women's self ratings and their ratings of their partners. The reported levels of minor and severe physical aggression in the homes of shelter women are in sharp contrast with the levels in the homes of community women, supporting the effectiveness of our selection procedures for securing samples of women with and without violence in their homes. The most severe tactics (beating, burning, using a gun) were virtually absent in the non-shelter comparison groups.
We examined Severe Physical Violence subscale scores for women's partners in relation to the presence or absence of threats to or actual harm and killing of pets. For both shelter groups (S-C and S-NC), we found the highest scores on partners' Severe Physical Violence in cases where these men had either threatened pets or had threatened and actually hurt or killed a pet.
We also correlated the severity of threats to pets and the severity of actual harm to pets with CTS subscale scores. For S-C women's self ratings on the CTS, we found the following significant correlations (there were no significant correlations in the S-NC, NS-C, and NS-NC groups):
|Threat Severity||Severity of Harm|
|Verbal Techniques||not significant||not significant|
|Verbal Aggression||-.52 (p=.03)||-.50 (p=.02)|
|Minor Physical Aggression||-.41 (p=.09)||-.57 (p=.004)|
|Severe Physical Aggression||-.71 (p=.002)||-.45 (p=.03)|
A similar analysis for women's CTS reports on their partners yielded significant correlations only for the NS-C group and only for threat severity (harming was too infrequent for analysis):
|Verbal Techniques||not significant|
|Verbal Aggression||-.85 (p=.03)|
|Minor Physical Aggression||-.75 (p=.09)|
|Severe Physical Aggression||not significant|
These data suggest that, in the community sample with children, a partner's lower use of verbal and minor physical aggression was associated with more severe threats toward pets. These results suggest the existence of different dynamics in the relation between partner violence and pet abuse for the shelter and community groups.
Finally, women were asked, "Are there any other pet or animal-related issues you'd like to discuss?" The shelter women invariably mentioned a litany of pet abuse including a partner prompting a dog fight, trying to hit a pet with a car, forcing his wife to have sex with a dog, threatening to drop an animal from the fourth floor of a building, and starving an animal to death. The community women, in contrast, mentioned positive features of pet ownership, including the role of pets in relieving stress, grief over a lost pet, a pet's role in helping a child overcome her fear of animals, and the calming effect of riding a horse.
The results of this study replicate and extend the results of an earlier study (Ascione, 1997) of 38 women who were battered and their reports of their partners' and children's maltreatment of pets. In the earlier study, 71% of the women with pets reported that their partner had either threatened or had actually hurt or killed one of their pets, and actual pet abuse was reported by 57% of the women. In the present study, 70.3 % of women in the shelter group reported either threats to or actual harm of pets with 54% reporting actual harm. Threats to pets were reported by 16.7% of non-shelter women and only 3.5% of these women reported actual harm of pets, a sharp contrast to the shelter group data. The advantages of the current study include the larger sample of shelter women (101 vs. 38), the inclusion of women from five different shelters instead of only one, and a comparison group of community women who did not report backgrounds of domestic violence. We now have more evidence that the results of the earlier study were not idiosyncratic and that families with domestic violence may experience significant levels of threats to and abuse of their pets.
Our incorporation of the Conflict Tactics Scale (CTS) in this study also serves to tie this research to the broader literature on family violence. We found that partners' use of extreme physical aggression toward their wives or female companions in the shelter sample was associated with the presence or absence of threats and harm to pets. However, it should be recalled that the severity of these threats and harm was inversely related, only in the shelter sample with children, to women's own use of more aggressive conflict tactics. One possible interpretation is that when women are themselves less violent, this may make a partner more prone to threaten or harm pets in a more serious manner, at least in families where children are present. Conversely, women who are more violent may have partners who are less prone to seriously threaten or hurt pets. For the community sample with children, we found that severity of threats (but not harm since the N was too small) was inversely related to partners' use of verbal and minor physical aggression (but not related to severe physical aggression). Thus, in generally nonviolent homes with children, partners' greater use of verbal and minor physical aggression is related to less severe threats toward pets. These potentially different dynamics, in the shelter and community samples, between CTS scores and threat/harm severity deserve further research attention.
Shelter women's emotional responses to their pets' victimization has not been explored in previous research. The intensely negative emotional responses these women had to threats to or harm of their pets suggests this is another form of trauma with which these women must cope. Both the women and their children acknowledged that young people in these families had witnessed pet abuse and the shelter children's emotional reactions to pets being hurt were also negative. One can speculate that the abuse of pets (and seeing how this affects their children) may prompt some women to decide to leave their violent partner yet, in a substantial minority of cases, women in our shelter sample said that concern for their pets' welfare actually delayed their leaving the home environment to seek shelter. This finding reinforces the importance of programs that provide women who are battered with a method of placing their pets in a safe haven when women decide to seek shelter. Other considerations for such programs are the generally lower level of veterinary care we found in our shelter sample (e.g., fewer animals having their vaccinations) and the apparently higher rate of pet "turnover" in these families as suggested by the number of pets owned in the past five years.
To further extend the findings from earlier research (Ascione, 1997), we documented higher rates and more severe levels of animal abuse by the children of shelter women. Although only one in ten of the S-C group women identified a target child as abusive of pets, 27.3% of S-C and 34.2% of S-NC women identified other children in their families who had abused pets (the rate found in Ascione  was 32%). These rates of child-perpetrated cruelty to animals are comparable to other studies of psychologically compromised children. The fact that more shelter than community group children scored in the clinical range on the checklist of child behavior problems also suggests that the symptoms of these children's psychological disturbance are not limited to animal abuse. Exposure to domestic violence is clearly a risk factor for children and it might be useful to consider exposure to pet abuse as a further compounding of risk.
And yet, despite the less than optimal rearing environments experienced by shelter group children, the vast majority of these children also provided care for pets and over half of the children reported that they had protected their pets from harm, often through active intervention that could have been dangerous to the children themselves. We know that children will sometimes try to intervene in altercations between their parents - we now know that they may also intervene to rescue their pets from harm. Future research should attend to the admixture of caring and cruelty present in children reared in violent homes. A better understanding of the factors that lead some children to vent their pain on those less vulnerable than themselves (for example, younger siblings or pets) and other children to cope with their pain by becoming more nurturing could inform prevention and intervention efforts. Given the discouraging national statistics on domestic violence, it is clear that many children and their pets would benefit from the application of this knowledge and understanding.
Although the samples of women we recruited for this study were comparable in age and in their volunteer status, we were discouraged by our failure to match the shelter and community samples more closely on demographic variables such as education and social status, ethnic diversity, and marital status. The unanticipated difficulty in recruiting and testing shelter women within our original time frame required us to begin testing community women before the demographics of the shelter sample became available. To address this limitation, we were fortunate to receive supplementary funding from Utah State University's Vice President for Research and will attempt to recruit additional samples of community women who do match the demographics we now have available. If we are successful, we will forward the results for these women to the Dodge Foundation as an addendum to the project report. It should be noted, however, that the demographics of our shelter sample did correspond well to the demographics of a state-wide population of spouse abuse victims studied in Utah (Thompson, 1994).
Future research should also include direct assessment (through interviews and use of the Child Behavior Checklist) of children from homes without domestic violence. We are currently examining the ethical issues involved in such research (since children would be asked about potentially disturbing incidents). We are doing this, in part, by becoming more familiar with the ethics of research on children's reports of community violence.
A final area of inquiry that should be noted is the need to interview domestic violence perpetrators about animal maltreatment. Partners may differ in their perceptions about the significance of cruelty to animals, especially its effects on children who witness it. Addressing animal abuse might also be a critical component of intervention programs for those who direct their violence toward intimate partners, their children, and their pets.
One of our primary goals is to prepare a report of this research suitable for submission to a scholarly journal in the area of interpersonal violence. We continue to review existing and newly published research in this area and monographs that deal with childhood and adolescent psychopathology. We are subjecting the data to appropriate statistical analyses and, once these are completed, will prepare a manuscript for submission to a journal. We will forward the Dodge Foundation a draft version of this document.
We will also prepare graphic representations of some of the results of this study for use in national presentations to animal welfare, child welfare, and domestic violence prevention programs. These will include a presentation in October at Michigan's annual state-wide conference for domestic violence specialists and submission for a proposed workshop at the 12th National Conference on Child Abuse and Neglect in 1998.
The sponsorship of the Geraldine R. Dodge Foundation will be acknowledged in each of these efforts at dissemination.
The attached listing of budget expenditures, provided by the Department of Psychology's administrative assistant, includes buyouts for the principal investigator and research assistants (Weber and Thompson). "Participants" refers to the women and children included in the samples and payments listed for Ctr. For Women - Provo, YCC of Ogden, Brigham YWCA, YWCA (Salt Lake City), and CAPSA (Logan) refer to the five cooperating shelters for women who are battered.
I thank Claudia Weber for her coordination and training efforts, often challenging to say the least, Teresa Thompson for her sensitive and professional assessment work, David Wood for his collaboration, and the directors and staff at each of the five shelters, without whose cooperation, there would have been little on which to report. Karen Ranson responded to the many phone inquiries we had about this project and facilitated the preparation of printed materials. Her professional secretarial contributions were invaluable.
A final and heartfelt thanks to the women and children who allowed us glimpses into their lives, their fears, their distresses, and their hopes to reside in landscapes of peace.
Are battered women in domestic violence shelters forced to chose between their personal safety and that of the pets they left behind when they fled? What policies and procedures do enlightened shelters employ to deal with the issue of pet abuse by batterers as a means of manipulation? What assistance can be provided? What are the psychological ramifications of pet abuse in a domestic violence context? Dr. Frank Ascione provides the answers in Safe Havens for Pets: Guidelines for Programs Sheltering Pets for Women Who Are Battered. This vital work is available FREEthanks to funding from the Geraldine R. Dodge Foundation. Click here for all the details. This book is must reading for every domestic violence worker, advocate, student, and supporter.
You can get more information on this subject from our Resources section Animal Abuse and Domestic Violence.
Achenbach, T. M. (1991). Child Behavior Checklist for Ages 4-8. University of Vermont, Burlington.
Ascione, F. R. (1997). Battered women's reports of their partners' and their children's cruelty to animals. Journal of Emotional Abuse.
Ascione, F. R., & Weber, C. (1995). Battered Partner Shelter Survey (BPSS!/Pet Maltreatment Survey. Logan: Department of Psychology, Utah State University.
Ascione, F. R., & Weber, C. (1995). Battered Partner Shelter Survey (BPSS!/Pet Maltreatment Survey (Mother/child version!. Logan: Department of Psychology, Utah State University.
Ascione, F. R., & Weber, C. (1995). Children's Observation and Experience with Their Pets (COEP!. Logan: Department of Psychology, Utah State University.
Ascione, F. R., & Weber, C. (1996). Families and Pets Survey 1996. Logan: Department of Psychology, Utah State University.
Straus, M. A. (1993). Conflict Tactics Scales forms. Durham: Family Research Laboratory, University of New Hampshire.
Thompson, K. D. (1994). Officially reported characteristics of spouse abuse victims seeking assistance in Utah. 1992. Unpublished master's thesis, Utah State University, Logan.