The animal--human bond and ethnic diversity.

Risley-Curtiss, Christina ; Holley, Lynn C. ; Wolf, Shapard 等

Keeping companion animals (that is, pets) is a universal cultural phenomenon (Brown, 002). In the United States, dogs or cats or both are found in at least one-third of households: Approximately 64 million households have a companion animal, including 77.7 million cats and 65 million dogs (American Pet Products Manufacturers Association, 2003).

An ever-increasing body of literature suggests that affectionate relationships with animal companions have health-enhancing effects on people and enrich their quality of life (see Lago, Delaney, Miller, & Grill, 1989; Netting, Wilson, & New, 1987; Sable, 1995). Research has also suggested that these relationships are complex and vary depending on a number of study population characteristics (Wilson & Netting, 1987). It stands to reason that ethnic and cultural differences may be among the factors that either enhance or inhibit such effects and that an understanding of the effect of ethnicity on relationships with companion animals is essential. Unfortunately, although there is a growing body of research on the animal--human bond, there appears to be little attention paid in the broad professional literature to ethnic and cultural beliefs, values, and practices regarding companion animals (Kaufman, 1999), and there is essentially none in the social work literature.

THE ANIMAL--HUMAN BOND AND SOCIAL WORK

The health and well-being of companion animals and humans have long been intertwined but have only recently been the subject of empirical study. Fortunately, knowledge of this relationship is growing, and there is much evidence attesting to the powerful connections between people and their animals, both positive and negative (Faver & Strand, 2003; Netting et al., 1987; Sable, 1995). Yet these animal--human connections have received virtually no attention from social work scholars, despite the fact that, for example, abuse and cruelty to companion animals in families is somewhat commonplace and has disturbing consequences for both humans and animals (see, for example, Deviney, Dickert, & Lockwood, 1998; Faver & Strand; Flynn, 2000). In addition, social work and other human service practitioners are increasingly including animals as adjuncts to treatment without social work research to inform and support them.

A recent Social Work Abstracts search of articles published since 1977 produced 25 references using the key word pet, with only 22 articles that addressed people's social relationships with animals. The key word animal drew 47 references, 33 of which referred to such things as animals in substance abuse or mental health research. Thus, in 27 years there have been no more than about 30 isolated articles on animal companions and humans in the social work literature.

Existing research has largely been conducted by those in psychology, psychiatry, and veterinary medicine and is often from an individual psychopathogical perspective. In contrast, social work is grounded in an ecological-systems perspective that requires one to view humans within the context of their environments and as constantly in reciprocal interaction with significant others (Hepworth, Rooney, & Larsen, 2002). Sixty-two percent of U.S. households have reported owning a companion animal (American Pet Products Manufacturers Association, 2003), which means that social workers are very likely to work with families with companion animals. Furthermore, because 68 percent of Americans consider companion animals as family (Brookman, 1999), the inclusion of companion animals in both practice and research should be a natural extension of social workers' work with humans and their challenges, coping mechanisms, and resilience factors. Recent related research, however, has suggested that this inclusion may not necessarily be happening. A study of 203 psychologists found although 94 percent believed animal abuse to be connected to other behavioral disturbances, only 14 percent assessed for such abuse (Nelson, 2002). A study of cross-reporting between child welfare workers and humane society workers found that a number of child welfare workers thought cross-reporting was unimportant and were resistant to including animal welfare in their assessments. These child welfare workers also underreported concern for animal well-being (Zilney & Zilney, 2005). Finally, an unpublished study by Risley-Curtiss (2004) found that only seven of 230 schools of social work that responded to a survey included much content on the human--animal bond in their courses, and what was offered most often dealt with animal-assisted therapy. Recognizing the potential impact of companion animals on people's lives could significantly affect social workers' abilities to help their clients. With appropriate knowledge and training, social work professionals are in a position to do much to enhance the lives of both people and their companion animals.

LITERATURE REVIEW

As described earlier, there is a dearth of literature on the effect of ethnic and cultural diversity on the animal--human bond in the United States. The 2002 edition of the U.S. Pet Ownership and Demographics Sourcebook, published by the American Veterinary Medical Association, does not include ethnicity in its companion animal owner demographics. In a study on the link between corporal punishment and children' cruelty to animals, Flynn (1999) incidentally noted that white people were somewhat more likely to commit animal abuse than people of color. Unfortunately, even with a sample about one-fifth African American, Flynn failed to take the opportunity to explore these ethnic differences in this and other articles drawing on the same data (for example, see Flynn, 2000). As noted by Kaufman (1999): "Rarely has animal protection in general, and the violence link specifically, been discussed in the context of a multicultural society" (p. 260).

Some ethnic differences have been found in companion animal ownership and attachment. Marx and colleagues (1988) examined demographics of companion animal ownership among U.S. adults ages 21 to 64 and found that white people were statistically more likely than people of color to have companion animals. Nonetheless, although race was identified as a significant factor, there was no discussion or examination of this finding. In a more in-depth examination, Siegel (1995) interviewed 877 urban non-Latina/Latino white, Latina/Latino, African American, Asian, and "other" adolescents between the ages of 12 and 17. The author found distinct racial or ethnic differences in companion animal ownership and attachment. White adolescents were most likely to have companion animals in their homes, followed by Latinos/Latinas and those of Asian descent. African Americans were least likely to have such animals in their homes. White youths also rated their companion animals as significantly more important to them than did Latina/ Latino youths, whereas African American and Asian youths reported intermediate levels of attachment but were not significantly different from any other group. Having sole responsibility for the care of a companion animal was also related to race and ethnicity, with African American youths most likely to have sole responsibility, followed by Latinas/Latinos, white, and Asian youths. Brown (2002) examined racial variations in pet attachment among 76 African American and 57 white veterinary students at one school. Among her findings was that more white students had companion animals (100 percent compared with 86 percent); they also had a larger number of such animals and were more likely to sleep with their companion animals than were African American students (70 percent compared with 53 percent). White students also had significantly higher scores on a pet attachment questionnaire than did African Americans. For example, there were differences in responses to such attachment statements as "No family is complete without a pet" and to questions about taking pets to visit friends and relatives. Brown concluded that how attachment to animals is shown may vary from culture to culture. She also noted that differences in housing, socioeconomic factors, and urban-rural background may aid in explaining the racial differences she found.

Johnson and Meadows (2002) examined 24 community-dwelling Latinas/Latinos age 50 and older regarding their relationships with their companion animals (dogs) and whether attachment to a companion animal was related to self-perceived health and functional ability. The majority of participants viewed their dogs as equals (54 percent) and as members of their family (79 percent). Sixty-seven percent stated that their dog was the reason they got up in the morning, and 62 percent stated that their dog comforted them. Although the study results did not support a relationship between animal companion relationship and participant's health or functional ability, the authors concluded that for these Latinos/Latinas, dogs were as important as they have been reported to be among white people and that dog companions were considered, among these elderly Latinas/Latinos to be valued members of their families.

Finally, the nonprofessional Best Friends magazine recently published the first in a planned series of articles on the lack of people of color in the animal welfare field (Richard, 2004). This first report is on African Americans and is anecdotal, but it makes an important observation regarding the lack of people of color in animal welfare work. Richard noted that "White people often believe that minority communities simply don't care about animals" (p. 14). This attitude is implied throughout the animal bond literature by the treatment of all companion animal owners as one and by the apparent lack of interest in studying populations of color (Kaufman, 1999).

Richard (2004) and Kaufman (1999) clearly indicated a need for more information on the possible effect of race and ethnicity in understanding people's relationships with companion animals. Do people from various racial and ethnic groups perceive that they derive similar benefits from their companion animals? Do companion animals play similar roles in the lives of people from various ethnic groups? Answers to these questions will allow social workers to conduct more complete, culturally relevant assessments of individual and family well-being; to better understand whether and when different types of animals have the potential to assist individuals and families in healing from trauma or in enriching their quality of life; and to assist social workers in group and other facilities in assessing the appropriateness of incorporating animal-assisted activities and therapy into their interventions. Social work, as a profession, needs to join the ranks of other disciplines in building its knowledge base about animal--human connections.

THE STUDY

The Setting

The survey was conducted by a survey research laboratory (SRL) in a large southwestern metropolitan county. It was the pilot for a proposed ongoing survey that will be carried out on an annual or semiannual basis. The survey was done by telephone between July 10, 2003, and August 6, 2003. A total of 587 interviews were completed.

Interviewers who carried out the survey were paid staff of the SRL with a minimum of 30 hours of training. Calls were made from the SRL phone bank, where shift supervisors monitored interviewers and provided feedback for each shift on standardized interviewing practices. Data were collected using a computer-assisted telephone interviewing system that administered the sample, ensured proper skip sequences, and allowed only valid punches.

Sampling

The sample was drawn as a list-assisted random digit dial, using the Genesys System. The Genesys database contains current listings of all active area codes and central offices in the United States, along with a database of all listed telephone numbers. The county area codes were specified. Within the 584 working exchanges, numbers were drawn from 100 blocks that had at least three listed numbers. This improves the efficiency of sampling by avoiding 100 blocks that have been assigned but not populated with residential phones, while still giving full probability of inclusion to new construction, people who have recently moved and are not listed, and people who have chosen not to be listed in telephone directories.

A total of 4,790 telephone numbers were called. Because the person in the household who answers the phone is not a random distribution, the most recent birthday selection method (Lavrakas, Stasny, & Harpuder, 2000) was used to select a random respondent. This technique has the advantage of not requiring a detailed enumeration at the beginning of the interview.

Each number was called a minimum of 10 times, filling a grid of morning, afternoon, evening, weekend, and weekday. If no contact was made after 10 calls (all results were continuous busy signals or rings with no answer), no further attempts were made. If any of the 10 attempts resulted in an indication that the number might have been a residential household, calls continued up to a maximum of 20 attempts. Once a residential household was reached and a respondent chosen, no artificial limit was set on the number of callbacks. A total of 20,037 phone calls were made over the course of the project to the 4,790 sample numbers.

The response rate was calculated using the American Association for Public Opinion Research (AAPOR) Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys (AAPOR, 2004). Using Response Rate 3, which estimates the proportion of cases of unknown eligibility that are actually eligible, the response rate was 33.4 percent. Although this response rate is lower than we might like, there is some recent evidence that lower rates may not harm data quality as much as feared. For example, Keeter and colleagues (2000) found very few differences in responses between a five-day phone survey with a 36 percent response rate and an eight-week study with a 60.6 percent response rate.

Measures

The data for this study were a subset of the data from a larger survey that consisted of 121 questions covering a variety of topics, including how participants liked their environment, their views on international adoption, and demographics such as gender, ethnicity, and income. The independent demographic variables used in our subset of data were racial/ethnic identities measured dichotomously. To avoid forcing multiethnic participants to choose only one racial or ethnic identity, race or ethnicity was measured in a series of separate categorical variables. Participants were asked if they were of Hispanic/Spanish origin (yes/no) and if they described themselves as white (yes/no), black/ African American (yes/no), American Indian (yes/ no), Asian (yes/no), or Pacific Islander (yes/no). Note that although the survey used the terms Hispanic/Spanish origin and American Indian, we are aware of important critiques of these terms and therefore prefer the terms Latina/Latino and indigenous peoples.

The dependent variables were selected from the 37 questions that asked specifically about pet ownership and care and beliefs regarding pets. Although we prefer the use of the more equitable term "companion animal," we recognize that common usage continues to be pet. Therefore, the word pet was used in the survey. Having a pet was measured as a dichotomous variable (yes/no). What kind of pet respondents had was measured in a series of dichotomous variables (yes/no) that covered eight types: cat, dog, fish, horse, rabbit, turtle, small rodent, and bird. All those having pets were asked whether they agreed or disagreed with the statements "It's okay to declaw cats" and "My pet is a member of my family." Those with dogs, cats, or both were also asked what their pet offered them, such as unconditional love (yes/no) or sense of personal safety (yes/no), as well as questions regarding care of their animal (for example, "Is your dog/cat spayed or neutered" [yes/no]). Finally all participants (with and without pets) were asked whether they agreed with belief statements such as "Animals do not feel pain" and "Animals can feel sadness, fear, happiness, and love."

Analyses

Data analyses rely on descriptive statistics such as frequencies, chi-square tests, and forward stepwise regression. All the dependent variables were categorical, with most, but not all, measured dichotomously. Ethnicity, the independent variable, was measured as multiple dichotomous variables (Hispanic/Spanish origin, white, black/African American, American Indian, Asian and Pacific Islander). Therefore we used chi-square to test for statistical significance between the multicategorical dependent variables and the dichotomous independent variables. Forward stepwise logistic regression was used to examine the relationship between various ethnic identities and dichotomous dependent variables. Forward stepwise regression starts with only the constant in the equation (that is, no variables). At each step, the variable with the smallest significance level for the score statistic below the chosen cutoff value (.05) is entered into the model.

RESULTS

Sample Description

The majority of the 587 participants identified themselves as white, not Hispanic (78.4 percent; n = 460) and female (60.1 percent; n = 353) (see Table 1, Total Study Sample column) with a mean age of 46 years (SD = 16.5; range = 18-92). Another 12.3 percent identified themselves as of Hispanic/Spanish origin. Of these 72 participants, 35 identified as Mexican American, 14 as Mexican, three as Puerto Rican, two as Cuban, and 13 as "other"; five did not categorize their Hispanic/ Spanish origin. Almost 75 percent of participants had some education beyond high school, with 40.4 percent having attended technical or trade school or some college; 21.1 percent having graduated from a four-year college, and 16.1 percent having attended or completed graduate school. Almost 63 percent of the sample were married or living together as married, and 61.9 percent were working for pay. The median annual household income ranged from $50,000 to less than $60,000, and almost 60 percent made between $30,000 and $79,999 annually. The average household size was 2.76 (SD = 1.4; range = 1-8), with 16 percent of participants riving alone and 54.8 percent living with one or two other people.

The study sample differs somewhat from the county population from which it was drawn. The study sample included more women than in the general population and fewer people between the ages of 18 and 24 (Table 1). However, 59.3 percent of the county population and 61.1 percent of the study sample were 25 to 54 years old. The study sample was similar to the county on income in the lower and highest income categories but differed somewhat in the $50,000-$99,000 category, with more of the study sample falling in this category. The study sample was also a better-educated group, with high school nongraduates underrepresented. Finally, Latinas/Latinos were underrepresented and indigenous peoples were overrepresented in the sample population.

Companion Animal Ownership and Care

Almost 63 percent (n = 368) of the participants had pets at the time of the survey. Dogs were the most common pet by a wide range, with 79.1 percent of participants with animals having dogs, 21.2 percent having fish (see Table 2). Slightly more than 13 percent had both cats and dogs.

The majority of participants with companion animals again were white, not Hispanic (77.2 percent), female (60.6 percent), married or living together as married (65.8 percent), and working for pay (67.3 percent). Similar to the entire sample, their mean age was 44 years old (range = 18-92; SD = 14.6), 77.6 percent had some kind of education beyond high school, and medium income was $50,000 to less than $60,000 annually.

In regard to the care of their animals, the most common residence of cats and dogs was both inside and outside (55.7 percent), followed by inside only/never going outside (35.2 percent). Slightly more than 85 percent (n = 300) said that their dog or cat was spayed or neutered, and 78.3 percent said they had a veterinarian for their dog or cat. The most common reason for not spaying or neutering the dog or cat was "planning to breed it" (18 percent) followed by "It is not old enough" (12 percent). Three participants said they could not afford it, and four did not believe that it is right to spay or neuter animals. Slightly less than 30 percent (n = 103) obtained their dog or cat from family or friends, 20.5 percent got their dog or cat from a breeder, and 21.3 percent got their dog or cat from some sort of shelter. About 11 percent said they obtained their animals from an "other" source. We are unsure what those other sources were. Finally, more than 97 percent of all pet owners agreed with the statement "My pet is a member of my family."

Ethnic Differences Regarding Companion Animals

Companion Animal Ownership. Participants who identified themselves as American Indian were most likely to have companion animals (73.5 percent; n = 25/34), followed by white people (65 percent; n = 319/491), people of Hispanic/Spanish heritage (56.9 percent; n = 41/72), African Americans (40.9 percent; n = 9/22), Pacific Islanders (40 percent; n = 2/5), and Asians (37.5 percent; n = 6/16). Logistic regression analysis was done with the six ethnicity variables entered as independent variables and pet ownership (yes/no) as the dependent variable. The overall regression model was significant, with a model chi-square of 8.07 (p = .004). Two variables met the entry criteria: describing oneself as white or as American Indian. Describing oneself as white entered first. The odds of those identifying themselves as white having pets increased by 48 percent compared with those not describing themselves as white. After controlling for those identifying as white, the odds of those identifying themselves as American Indian having pets increased by 58 percent compared with those who did not so identify themselves.

To explore whether there is a significant relationship between specific types of companion animals and ethnic identity or identities, we ran a series of eight logistic regressions. For each equation, all six race and ethnicity variables were entered as independent variables and one type of companion animal was entered as the dichotomous dependent variable (that is, cat, dog, bird, rabbit, turtle, fish, horse, or small animals such as mice, rats, and hamsters) .These analyses revealed that identifying oneself as being of Hispanic/Spanish origin significantly reduced the odds of having cats by 59 percent (model [chi square] = 14.261,p = .01) and of having birds by 68 percent (model [chi square] = 7.23, p = .007). Describing oneself as American Indian increased the odds of having fish by 65 percent (model [chi square] = 5.34, p = .02). No other significant relationships were found between racial or ethnic group membership and type of companion animals.

Care of Companion Animals. Four questions were asked of those with a cat, dog, or both regarding the care of these animals: (1) whether their cat or dog was spayed or neutered; (2) what their reasons were for not spaying or neutering; (3) whether their dog or cat resided inside only, inside and outside, or outside only; and (4) whether they had a veterinarian for their dog or cat. These questions were used as the dependent variables in a series of logistic regressions with racial and ethnic identities as the independent variables. Those who identified as being of Hispanic/Spanish origin were 3.41 times more likely not to have their animal spayed or neutered than those who did not identify as of Hispanic/ Spanish origin (model [chi square] = 8.77, p = .003). We also used the chi-square test of independence for two multicategorical dependent variables: reasons for not spaying or neutering their cat or dog and residence of dog or cat (inside only, inside and outside, outside only). Given the high frequency of cells with expected counts of less than five, we were unable to test for statistical significance between ethnic identity and reasons given for not spaying or neutering their dog or cat. When we collapsed the residence variable to two categories (inside and outside versus outside only), the nonparametric Fisher's Exact Test was significant for identifying as white (p = .04). However when identifying as white was entered into a logistic regression along with the other ethnicities as independent variables, the model was not significant. Finally, cat and dog owners were asked whether they had a veterinarian for their pet. Hispanic/Spanish origin was the only ethnicity to enter the model with having a veterinarian for their pet as the dependent variable (model [chi square] = 6.49,p = .01). Those identifying themselves as of Hispanic/Spanish origin were 73 percent more likely not to have a veterinarian for their pet than were those who did not identify themselves as of Hispanic/Spanish origin.

Relationships and Beliefs. Cat and dog owners were also asked a series of questions regarding their relationships with their companion animals, including what their animal offered them. Specifically they were asked if their pets offered them emotional support, unconditional love, companionship, a sense of personal safety, income from breeding, and a chance to teach their child responsibility. In addition, all pet owners were asked if they agreed to such statements as "My pet is a member of my family" (agree/disagree), and "It is okay to declaw a cat" (agree/disagree). These variables were used as dependent variables with all ethnic identities as independent variables in logistic regression analyses. No statistically significant differences were found with regard to racial and ethnic identity and emotional support, unconditional love from pet, companionship from pet, and pets are family members. Those identifying themselves as Hispanic/ Spanish origin were 55 percent more likely to get a sense of personal safety from their pet than were those not identifying themselves as Hispanic/ Spanish origin (model [chi square] = 4.36, p = .04). Those identifying themselves as Asian were 93 percent more likely to say they got income from breeding than were those not identifying as Asian (model [chi square] = 18.91, p = .000). It should be noted, however, that there was a very small number of Asian Americans with pets (n = 6), and only two (33 percent) said they got income from breeding. Those identifying themselves as white were 2.44 times more likely not to feel that their pets taught their children responsibility than were those not identifying themselves as white (model [chi-square] = 5.10, p = .02). Finally, the odds of those identifying themselves as white agreeing that it is okay to declaw cats increased by 64 percent when compared with those not identifying themselves as white (model [chi square] 8.56, p = .003).

Participants were asked whether they agreed with another series of statements, whether they had pets or not. These included "Pets should be spayed/neutered early in life," "It's okay to hit a pet if it misbehaves," "Animals do not feel pain," "Animals can feel sadness, fear, happiness, and love," and "There are too many unwanted animals in the world. "All were measured dichotomously as agree/disagree. Frequencies by ethnicity are presented in Table 3. Statistical significance was found only in the logistic regression model for "pets should be spayed/ neutered early in life" and for whether animals feel pain. Those who identified themselves as white were 0.59 times more likely to agree that animals should be spayed or neutered early in life than those not identifying themselves as white (model [chi-square] = 7.57,p = .006). Those who identified themselves as Asian were 81 percent more likely to agree that animals do not feel pain (model [chi square] = 4.58, p = .03). In addition, although the overall model was not statistically significant, identifying oneself as Asian was associated with disagreeing that there are too many unwanted animals in the world (p = .05).

Limitations of the Study

Before we discuss the implications of this study, several limitations are important to consider. First, although a random sampling technique was used, people of color were underrepresented in the sample in comparison to their representation in the population from which they were drawn. This underrepresentation limits the study's generalizability. Furthermore, the population from which the sample was drawn is one with relatively few people of Asian, African American, or Pacific Islander descent. This limitation prevented detection of significant differences that may actually exist among these racial and ethnic groups. Future studies might use a stratified random sampling technique to address this problem. Second, because of the exploratory nature of the study and the need to limit the number of questions asked, relatively few questions regarding human-companion animal attachment were asked, and no questions were asked about why participants did not have companion animals. Future researchers might benefit from asking additional questions about the animal-human relationship (for example, "Does your pet sleep in your bed with you?") and about factors that influence the decision to have or not have companion animals.

Other research has suggested that care practices and beliefs may vary depending on an individual's attachment to a companion animal (Wilson & Netting, 1987). Finally, the use of stepwise regression has limitations, in that a model is selected to fit a particular sample. Thus, there is no assurance that the same model would be selected if another sample from the same population were selected. In addition, the model always fits the sample better than the population from which it was selected (Norusis, 1990). One, therefore, must use caution in interpreting the results, especially with regard to other populations.

DISCUSSION

It is interesting to note that the percentage of pet owners in our study is almost exactly the same (62 percent versus 63 percent) as that found in the 2002 American Pet Products Manufacturers Association survey (American Pet Products Manufacturers Association, 2003). In addition, this study's findings offer support that there are some ethnic differences in whether people have animal companions and in the care of and beliefs regarding companion animals. A higher percentage of indigenous people indicated that they have pets; those who identified as white were next most likely to have pets. This finding is particularly important because no earlier studies have included indigenous people in their study samples.

Furthermore, members of different racial or ethnic groups who do have pets may prefer the companionship of different types of animals. In this sample, those who identified as being Latina/Latino were significantly less likely than others to have cats or birds, whereas those identifying as indigenous were more likely to have fish. Although no other statistically significant differences were found in types of companion animals owned by members of various ethnic groups, it is possible that a sample that includes larger numbers of people who identify as African American, Asian, and Pacific Islander may reveal other differences among groups.

Perhaps the most important finding is that among study participants who had cats, dogs, or both, there were no significant racial or ethnic differences in whether participants reported that they receive emotional support, unconditional love, and companionship from their pets. Furthermore, regardless of racial or ethnic group membership, almost all companion animal owners described their animals as being family members. This finding is consistent with earlier studies (Brookman, 1999) and reinforces the importance of social workers' considering clients' relationships with their companion animals in their work with individuals and families (see Implications for Social Work Practice section, below). Of course, companionship, emotional support, and so forth may mean different things to different people, so further research is needed to better understand these findings (see Implications for Social Work Research section).

The finding that Latinas/Latinos report that their dogs or cats give them a sense of personal safety is also critical to consider. Because members of this group are less likely to have cats, it may be assumed that they considered their dogs as providing this sense of safety. Although further research is certainly needed to understand this finding, it is not surprising that a social group that experiences discrimination and other forms of oppression might gain a sense of safety from having a dog companion.

This study found that those who identified themselves as white were significantly different from others on three outcome variables. White people were significantly more likely than others to say that pets should be spayed or neutered early in life and that it is okay to declaw cats and least likely to say that having a cat or dog teaches children to be responsible. Further research is needed to learn whether this first difference is related to socioeconomic status. The first difference suggests a concern for the nonhuman animal world, whereas the second and third differences appear to be more focused on humans' needs or preferences. Further research is needed to better understand these apparently complicated attitudes about companion animals.

IMPLICATIONS FOR SOCIAL WORK RESEARCH

Findings from this exploratory study have several implications for future research. First and most critical, the finding that the vast majority of participants with companion animals see them as family members strongly indicates that social work researchers who are studying individual and families must include, at a minimum, the presence of companion animals--and their meaning to human participants from various ethnic, gender, age, and other social groups--in their research. Research from other disciplines has already suggested that animals can play an important role in violent families experiencing domestic violence or child abuse (for example, see Flynn 1999, 2000) in the development of children (Melson, 2001), and in enhancing the fives of elderly people (Baun & McCabe, 2000). By including animals in their studies, social work researchers can add to this important, growing body of knowledge and assist social work practitioners in implementing a more inclusive ecological approach to work with clients.

In relation to ethnicity, additional quantitative studies need to be conducted with samples that include larger numbers of people from various racial and ethnic groups, including Latinas/Latinos who are not of Mexican descent. Studies might also include measures of the strength of ethnic identity (for example, see Cuellar, Arnold, & Gonzalez, 1995; Klonoff & Landrine, 1999; Resnicow, Soler, Braithwaite, Selassie, & Smith, 1997; Roberts et al., 1999). Studies in other areas have found that strength of ethnic identity, rather than simply ethnic group membership, may affect attitudes and behaviors (for example, see Lorenzo-Hernandez & Ouellette, 1998; Phinney, Cantu, & Kurtz, 1997). Such studies also need to examine whether there are differences in attitudes toward and relationships with animals among people from various geographic regions, ages, genders, social classes, and urban, rural, and suburban residences and those with different amounts of leisure time and varying personal experience with companion animals. Future studies might include additional questions that will allow researchers to better understand people's relationships with their companion animals. For example, are there ethnic differences in whether it is appropriate to allow one's dog to sleep in one's bed? Are there differences in how much time one spends with one's cats? This study found that people obtain unconditional love, emotional support, and companionship from their cats and dogs. Additional research is needed to understand human relationships with other types of companion animals. How would people from various ethnic, gender, and age groups rate the importance of the benefits of having different types of companion animals?

Qualitative studies are needed to assist in understanding what these relationships mean to people. For example, when people say that they get emotional support from their companion animals, what do they mean? When do they turn to their animal for this emotional support? Are there instances in which this emotional support is more important than what one might obtain from a human companion--or from a social worker? Can an animal who is not a long-term part of one's family of origin or choice (for example, a horse in an outdoor program for youths; a cat in a residential care facility) provide the same level of companionship, unconditional love, and emotional support as a family's companion animal? Does such an animal become part of an individual's family, as study participants described? Are there ethnic, age, or other social group differences and similarities in how people experience these relationships?

Research is also needed to learn whether individuals living in stressful environments (for example, higher crime areas, violent or neglectful families) have different relationships with their companion animals than do others. It is possible that people in such environments may feel that the responsibilities of caring for an animal simply add stress to their lives. Alternatively, such people may benefit from the companionship, emotional support, and unconditional love of a companion animal.

Unfortunately, it makes sense that some people who consider their companion animals to be members of their families might abuse animals just as they might abuse humans. More studies of the factors related to animal abuse--particularly studies that consider whether there is a relationship between animal abuse and human abuse are needed to help us better understand both of these social problems.

Studies also need to explore the perspectives of the sizable minority of people who do not have companion animals. Is not having companion animals related to living situation, income, formal educational level, attitudes about animals, or something else?

Finally, available research suggests that many social workers may not be including attention to animal--human relationships in their practices. Additional studies are needed to understand if some social workers are doing so and ways in which practice that includes attention to such relationships affects the lives of clients.

IMPLICATIONS FOR SOCIAL WORK PRACTICE

The finding that the vast majority of dog and cat owners in this study said that their animal companions provide them with emotional support, unconditional love, and companionship and that they consider their animals to be family members has tremendous implications for social work practice, particularly in light of evidence suggesting that attention to animals may currently be lacking (for example, see Risley-Curtiss, 2004; Zilney & Zilney, 2005). First, these findings support the practice of asking about the presence of animals and including attention to individuals' relationships with their companion animals, particularly cats and dogs, in biopsychosocial assessments. If animals are present and not included, essential elements of individuals' interactions with their environments will be missing, and the interventions that follow may fail to draw on critical resources for clients' health and well--being. In addition, if an individual is strongly attached to her or his companion animal, an adequate intervention may need to include providing resources for care of a companion animal (for example, food; veterinary care, including vaccinations; boarding services during hospitalization or other out-of-home care of the person). Although such components are rarely considered "social work interventions," they may indeed be important for clients' well-being.

Findings also have implications for work with specific racial and ethnic groups. For example, social workers who want to include animals in their group care settings that serve Latinas/Latinos (at least those who are of Mexican descent) may consider selecting dogs, rather than cats or birds, for residence in their facilities. Findings also suggest that facilities serving indigenous and white populations, in particular, might improve their services by including animals in their programs.

Finally, findings suggest that adult group homes, shelters, and other care facilities might consider the feasibility of allowing people to bring their companion animals with them. Of course, there are many difficult issues involved in such an arrangement (such as the potential for an animal to harm another person or animal, or vice versa). Instead, social workers might ensure that animals are well cared for when apart from their human companions and arrange for visitation opportunities. This could decrease individuals' worries about the well-being of their animal companions, while continuing to provide benefits resulting from the emotional support, companionship, and perceived unconditional love of a companion animal.

Given the findings of this and other studies regarding the importance of companion animals, social work researchers and practitioners must pay attention to the presence of animals in the lives of their clients. To give companion animals short shrift means we are missing a potentially vital connection and are paying only lip service to our claims of an ecological approach.

Original manuscript received March 19, 2004 Final revision received March 18, 2005 Accepted June 1, 2005

REFERENCES

American Association for Public Opinion Research. (2004). Standard definitions: Final dispositions of case codes and outcome rates for surveys (3rd ed.). Lexena, KS: Author.

American Pet Products Manufacturers Association. (2003, April 14). Leading trade organization releases study finding pets owned by more than half of all U.S. households. Retrieved January 29, 2004, from www.appma.org/ press/press releases/2003/nr_04-14-03.asp

American Veterinary Medical Association. (2002). U.S. pet ownership and demographics sourcebook. Schaumburg, IL: Author.

Baun, M. M., & McCabe, B.W. (2000).The role animals play in enhancing quality of life for the elderly. In A. Fine (Ed.), Handbook of animal-assisted therapy (pp. 237-251). New York: Academic Press.

Brookman, F. (1999, October 25).The things people do for pets! Discount Store News. Retrieved January 29, 2004, from www.findarticles.com/cf_dls/_m3092/ 20_38/57443535/p1/article.jhtml

Brown, S. E. (2002). Ethnic variations in pet attachment among students at an American school of veterinary medicine. Society and Animals, 10, 455-456.

Cuellar, I., Arnold, B., & Gonzalez, G. (1995). Acculturation Rating Scale for Mexican--Americans--II: A revision of the original ARSMA scale. Hispanic Journal of Behavioral Sciences, 17, 275-304.

Deviney, E., Dickert, J., & Lockwood, R. (1998). The care of pets within child abusing families. In R. Lockwood & F. Ascione (Eds.), Cruelty to animals and interpersonal violence: Readings in research and application (pp. 305-313). West Lafayette, IN: Purdue University Press.

Faver, C. A., & Strand, E. B. (2003). Domestic violence and animal cruelty: Untangling the web of abuse. Journal of Social Work Education, 39, 237-253.

Flynn, C. P. (1999). Exploring the link between corporal punishment and children's cruelty to animals. Journal of Marriage and the Family, 61, 971-981.

Flynn, C. P. (2000). Why family professionals can no longer ignore violence towards animals. Family Relations, 49, 87-95.

Hepworth, D. H., Rooney, R. H., & Larsen, J.A. (2002). Direct social work practice (6th ed.). Pacific Grove, CA: Brooks/Cole.

Johnson, R.A., & Meadows, R. L. (2002). Older Latinos, pets and health. Western Journal of Nursing Research, 24, 609-620.

Kaufman, M. E. (1999). The relevance of cultural competence to the link between violence to animals and people. In F. R. Ascione & P. Arkow (Eds.), Child abuse, domestic violence, and animal abuse (pp. 260-270). West Lafayette, IN: Purdue University Press.

Keeter, S., Miller, C., Kohut, A., Groves, R.M., & Presser, S. (2000). Consequences of reducing nonresponse in a national telephone survey. Public Opinion Quarterly, 64, 125-148.

Klonoff, E. A., & Landrine, H. (1999). Acculturation and cigarette smoking among African Americans: Replication and implications for prevention and cessation programs. Journal of Behavioral Medicine, 22, 195-204.

Lago, D., Delaney, M., Miller, M., & Grill, C. (1989). Companion animals, attitudes towards pets, and health outcomes among the elderly: A long-term follow-up. Anthozoos, 3, 25-34.

Lavrakas, P. J., Stasny, E. A., & Harpuder, B. (2000, May 18-21). A further investigation of the last-birthday respondent selection method and within unit coverage error. Paper presented at the 55th Annual American Association for Public Opinion Research Convention, Portland, OR.

Lorenzo-Hernandez, J., & Ouellette, S. C. (1998). Ethnic identity, self esteem, and values in Dominicans, Puerto Ricans, and African Americans. Journal of Applied Social Psychology, 28, 2007-2024.

Marx, M. B., Stallones, L., Garrity, T. F., & Johnson, T. P. (1988). Demographics of pet ownership among U.S. adults 21 to 64 years of age. Anthrozoos, 2, 33-37.

Melson, G. F. (2001). Why the wild things are. Cambridge, MA: Harvard University Press.

Nelson, P. (2002). The survey of psychologists' attitudes, opinions, and clinical experiences with animal abuse (Doctoral dissertation, The Wright Institute, 2002). Dissertation Abstracts International, 63, 1040

Netting, F. E., Wilson, C. C., & New, J. C. (1987). The human-animal bond: Implications for practice. Social Work, 32, 60-64.

Norusis, M. J. (1990). SPSS/PC+ advanced statistics 4. O. Chicago: SPSS.

Phinney, J. S., Cantu, C. L., & Kurtz, D.A. (1997). Ethnic and American identity as predictors of self-esteem among African American, Latino, and White adolescents. Journal of Youth and Adolescence, 26, 165-185.

Resnicow, K., Soler, R. E., Braithwaite, R. L., Selassie, M. B., & Smith, M. (1997). Development of a racial and ethnic identity scale for African American adolescents: The survey of black life. Journal of Black Psychology, 25, 171-188.

Richard, J. (2004, January/February). Minority report: part one: African Americans. Best Friends, pp. 14-16.

Risley-Curtiss, C. (2004). [The animal--human bond and schools of social work]. Unpublished raw data.

Roberts, R. E., Phinney, J. S., Masse, L. C., Chen, Y. R., Roberts, C. R., & Romero, A. (1999).The structure of ethnic identity of young adolescents from diverse ethnocultural groups. Journal of Early Adolescence, 19, 301-322.

Sable, P. (1995). Pets, attachment, and well-being across the life cycle. Social Work, 40, 334-341.

Siegel, J. M. (1995). Pet ownership and the importance of pets among adolescents. Anthrozoos, 8, 217-223.

Wilson, C.C., & Netting, F. E. (1987). New directions: Challenges for human-animal bond research and the elderly. Journal of Applied Gerontology, 6, 189-200.

Zilney, L. A., & Zilney, M. (2005). Reunification of child and animal welfare agencies: Cross-reporting of abuse in Wellington County, Ontario. Child Welfare, 84, 47-66.

Christina Risley-Curtiss, PhD, MSSW, is associate professor, School of Social Work, Arizona State University, Tempe, AZ 85287-1802; e-mail: risley.curtiss@asu.edu. Lynn C. Holley, PhD, ACSW, is assistant professor, School of Social Work, Arizona State University. Shapard Wolf, MEd, is director and technology support analyst, principal, School of Life Sciences, Arizona State University. This study was part of a larger study conducted through the Arizona State University Survey Research Laboratory of which Mr. Wolf was, at the time, the director.

Table 1: Demographic Characteristics of County Population,
Sample Participants, and Pet Owners, 78 Years and Older

 County Total
 Population Study Sample
 (N = 2,244,146) (N = 587)

Characteristic % n % n

 n = 2,244,146 n = 585

 Female 50.5 1,132,745 60.1 353
 Male 49.5 1,111,401 39.7 232

Race or ethnicity n = 2,244,146 n = 587

 White, not Hispanic 71.0 1,593,872 78.4 460
 Hispanic, any race 20.8 466,312 12.3 72
 American Indian,
 not Hispanic 1.4 30,317 4.9 29
 Other race
 and mixed race 6.9 153,645 8.5 50

Age n = 2,244,146 n = 572

 18-24 14.0 313,365 9.3 53
 25-34 21.8 488,329 19.0 109
 35-44 21.2 475,907 21.5 123
 45-54 16.3 366,464 20.5 117
 55-64 10.7 241,102 13.1 75
 65 & over 16.0 358,979 16.6 95

Household income n = 1,133,048 n = 514

 Less than $10,000 6.8 77,072 6.2 32
 $10,000-$49,999 48.0 544,576 42.2 217
 $50,000-$99,999 31.9 361,254 40.3 207
 $100,000 or more 13.3 150,146 11.3 58

Educational attainment n = 2,246,838 n = 584

 Did not complete
 high school 19.5 438,609 3.9 23
 High school
 diploma or GED 23.5 528,340 18.5 108
 Some college/
 technical school/
 associate's degree 33.8 759,548 40.4 236

 Study Sample:
 Only Those
 with Pets
 (N = 368)

Characteristic % n

 n = 366
 Female 60.6 223
 Male 39.1 143

Race or ethnicity n = 368
 White, not Hispanic 82.1 302
 Hispanic, any race 11.1 41
 American Indian,
 not Hispanic 6.3 23
 Other race
 and mixed race 5.4 20

Age n = 362
 18-24 8.3 30
 25-34 20.7 75
 35-44 23.2 84
 45-54 24.6 89
 55-64 14.1 51
 65 & over 9.1 33

Household income n = 329
 Less than $10,000 3.0 10
 $10,000-$49,999 40.7 1.34
 $50,000-$99,999 43.2 142
 $100,000 or more 13.1 43

Educational attainment n = 367
 Did not complete
 high school 3.3 12
 High school
 diploma or GED 19.1 70
 Some college/
 technical school/
 associate's degree 40.6 149

Note: GED = general equivalency diploma.
The ns for race or ethnicity add up to more
than 587 because participants could identify
themselves as more than one race or ethnicity.

Table 2: Types and Frequencies of Companion
Animals by Ethnicity (n = 368)

 Pet Owners Hispanic/ White,
 with This Spanish not American
 Animal Origin Hispanic Indian
Type of
Animal n % % (n = 41) (n = 302) (n = 25)

Cat 61 16.6 5 55 3
Dog 214 58.2 29 180 14
Bird 46 12.5 11 37 5
Rabbit 3 0.8 0 3 1
Fish 78 21.2 12 65 10
Small animals
 (rats, etc.) 25 6.8 2 22 4
Horse 15 4.1 1 13 2
Turtle 14 3.8 4 13 1

 African Pacific
 American Asian Islander
Type of
Animal (n = 9) (n = 6) (n = 2)

Cat 0 2 0
Dog 9 4 2
Bird 0 2 0
Rabbit 0 0 0
Fish 0 1 0
Small animals
 (rats, etc.) 0 0 0
Horse 0 0 0
Turtle 0 0 0

Note: The ns for race or ethnicity add up to more than
369 because participants could identify themselves as
more than one race or ethnicity. Percentages add up to
more than 100 because participants could have more than
one type of pet.

Table 3: Beliefs Regarding Companion Animals
by Ethnicity (n = 587)

 Hispanic/ White,
 Spanish not African
 Origin Hispanic American
Belief Variable (n = 72) (n = 491) (n = 22)

Pets should be spayed or
neutered early in life.
 Agree 55 419 14
 Disagree 13 46 6
 Don't know (a) 4 26 2
It's okay to hit a
pet if it misbehaves.
 Agree 14 86 7
 Disagree 58 390 14
 Don't know (a) 0 15 1
Animals do

not feel pain.
 Agree 5 21 1
 Disagree 66 469 21
 Don't know (a) 1 1 0
Animals can feel fear,
sadness, happiness,
and love.
 Agree 70 484 20
 Disagree 0 4 0
 Don't know (a) 2 3 2
There are too many
unwanted animals
in the world.
 Agree 64 447 20
 Disagree 6 29 0
 Don't know (a) 2 15 2

 American Pacific
 Indian Asian Islander
Belief Variable (n = 34) (n = 16) (n = 5)

Pets should be spayed or
neutered early in life.
 Agree 27 10 4
 Disagree 6 1 1
 Don't know (a) 1 5 0
It's okay to hit a
pet if it misbehaves.
 Agree 4 3 1
 Disagree 29 11 4
 Don't know (a) 1 2 0
Animals do
not feel pain.
 Agree 3 3 1
 Disagree 31 13 4
 Don't know (a) 0 0 0
Animals can feel fear,
sadness, happiness,
and love.
 Agree 33 16 5
 Disagree 0 0 0
 Don't know (a) 1 0 0
There are too many
unwanted animals
in the world.
 Agree 29 13 5
 Disagree 4 3 0
 Don't know (a) 1 0 0

Note: The ns for ethnicity add up to more than 368
because participants could identify themselves as
more than one ethnicity. "Don't know" was coded
"don't know/refuse to answer."