[Reprinted with permission]

  In the Spring of 2003, I moved from Virginia to Texas to begin work as a tenure-track faculty member at Texas Tech University School of Law. I brought my two dogs with me: Saffy (a 4-year-old mixed breed whose parents were a fluffy red Chow Chow and a big black Labrador Retriever) and Semona (a 2-year-old Rottweiler). Neither Semona nor Saffy has ever bitten anyone. Neither has shown any aggressive tendencies. Both are extremely playful and friendly animals.

  After I placed a bid on a house in Lubbock, Texas, I began the search for homeowners' insurance-a process which I thought would be straightforward and easy. Much to my surprise, dozens of insurance companies denied my application outright. The reason? Semona is a Rottweiler and Saffy is half-Chow. Rottweilers and Chow Chows are on the "blacklist" of dog breeds. Some insurance companies believe they, along with Pit Bulls, Huskies, Doberman Pinchers, and other specified breeds, are more likely to bite humans and, in turn, cause liability claims to be brought against their owners. Even mixed breeds, like my half-Chow Saffy, are blacklisted. This practice is known by many dog owners as "breed discrimination."

  Thankfully, the story ended happily for my dogs and me. After weeks of calling nearly every insurance agent in Lubbock, I was able to obtain insurance through the Texas Farm Bureau, an organization that advocates for farmers and farming issues. [FN1] Had it not been for the Farm Bureau, I would have found myself on the horns of a horrible dilemma: whether to buy a home or give up my dogs. Anyone who knows me can confirm that this dilemma would have been easy to resolve; I would have chosen my furry family members over home ownership. [FN2] Sadly, however, many *4 Americans are finding themselves in similar positions and are opting to give up their dogs to animal shelters. [FN3]

  Breed discrimination by insurance companies is on the rise in the United States. Insurers are refusing to write homeowners' policies for people who own breeds that the insurance industry considers to be dangerous. Their decisions are based solely on the breed of the animal, notthe individual characteristics of the particular dog. Dog bites are certainly a public health concern. However, the insurance industry's approach to the problem is based on faulty assumptions and improper use of dog bite statistics. The insurance industry has prejudged entire breeds of dogs as being "too risky," instead of taking a more reasonable dog-to-dog approach to risk assessment.

  Major veterinary and breed registry organizations have strongly opposed breed discrimination in insurance. Authors of scientific studies on dog bites have even argued against the use of their data to support breed-based decision-making by insurers and legislatures. Dog owners across the country have spoken out about the horrible choice they have been forced to make between obtaining insurance and keeping their dogs.

  There has existed a historic tension between risk classification and social policy. Classification and insurability decisions are usually "actuarially justified"-that is, the insurance company has identified a statistical correlation between a characteristic and increased risk. Actuarial justification is frequently cited by insurers as a reason to avoid social regulation. Insurers exist to make a profit for their shareholders. They do so by minimizing risk which, in turn, minimizes claims paid out.

  Actuarial justification is only the first step in determining the social propriety of a proposed underwriting mechanism. Social utility of the risky conduct must also be considered. Statutes across the United States are replete with examples of legislatures overruling actuarially justified practices in favor of competing social policies. "Red-lining" is a classic example. Actuaries identified statistical correlations between living in certain neighborhoods and increased risk for claims against homeowners' policies. As a result, insurance companies began to refuse to write policies in these high-risk neighborhoods. The neighborhoods in question were often economically depressed and occupied by members of racial or ethnic minorities. Legislatures and courts stepped in to prohibit red-lining, despite the actuarial justification for the practice. [FN4]

  *5 Breed discrimination is a different animal altogether. Even without considering the high social utility of pet ownership, insurers have been unable to demonstrate an actuarial justification for discriminating based on breed. As the multidisciplinary Task Force on Canine Aggression and Human-Canine Interactions concluded, "[d]og bite statistics are not really statistics, and they do not give an accurate picture of dogs that bite." [FN5] The popular notion that Pit Bulls and Rottweilers are inherently more likely to bite is simply not supported by the available statistics.

  When the social utility of pets is added to the equation, breed discrimination becomes even more unreasonable. Dogs and other domesticated animals provide immeasurable joy and happiness to the families that own them. Even some components of the legal system itself have evolved to recognize pets as being more than mere chattel. [FN6] In addition, the failure to obtain homeowners' insurance is a death knell for homeownership-no insurance, no mortgage; no mortgage, no house.

  My argument is quite simple: Decisions regarding the provision, rating, termination, or renewal of a homeowners' insurance policy should not be based on ownership or possession of a particular breed of dog unless there is evidence of dog-specific risk. Insurers would concededly be actuarially justified in charging higher premiums or declining coverage for people who own dogs that have unjustly bitten in the past. After all, the best predictor of future behavior is past behavior. Breed discrimination, as it currently stands, is not actuarially justified because scientists have not been able to accurately determine whether certain breeds are inherently more dangerous, or instead whether a breed's high population is making it appear that the breed is more dangerous.

  The consequences of breed discrimination could not be greater. Homeowners' insurance is the gatekeeper to homeownership. Without homeowners' insurance, a buyer cannot get a mortgage. For most Americans, if a person cannot obtain a mortgage, he cannot buy a home.

  In Part I of this article, I will give an overview of the problem: dog breed discrimination by insurers, as well as a related problem of breed-specific legislation by some states. In Part II, I will analyze the major scientific studies on dog bites, showing that no one has adequately proven *6 that some breeds are more inherently dangerous than others. In Part III, I will show that breed discrimination and breed-specific legislation are opposed by most veterinary and animal groups. Part IV will demonstrate that insurers have been ignoring the unique and special role that pets play in millions of American homes. I draw upon not only the profoundly personal arguments advanced by myself and others, but also the way in which the law itself is evolving by recognizing pets as more than mere property. Part V shows how the insurance industry is a highly regulated industry which subjects itself to legislative control where, as here, the public is being harmed by underwriting decisions not driven by actuarial justification. I will also offer a number of alternatives to breed discrimination.

  Breed discrimination in insurance is a recent phenomenon that was preceded by the enactment of "breed-specific legislation" ("BSL") by some state legislatures and municipalities. Both breed discrimination and BSL are a perceived response to highly publicized attacks by certain breeds, particularly Pit Bulls.

A. Highly Publicized Attacks by Pit Bulls

  In the 1980s, there were a number of high-profile attacks on humans by Pit Bulls. These attacks led to a near-hysterical reaction by members of the communities that were affected by the attacks and by the legislators who represented them.

  In March of 1984, Pit Bulls attacked Angie Hands, a 9-year-old girl in Tijeras, New Mexico. [FN7] The dogs bit her right leg to the bone, ripped flesh from her arms, and tore her ear in half. [FN8] The child survived but had to undergo years of reconstructive surgery. [FN9] She had been attacked by her uncle's four Pit Bulls in between her bus stop and her home. [FN10] The small community of Tijeras-located outside of Albuquerque-responded with an outright ban on Pit Bull ownership. [FN11] Dog owners challenged the law in *7 court, but the law was upheld as a constitutional exercise of the town's police power. [FN12]

  The attack on Angie Hands followed a number of other Pit Bull attacks around the country. A 4-year-old girl was killed in Oregon City when she fell into a yard where a Pit Bull was chained. [FN13] Two Pit Bulls mutilated their owner in Edgemere, Maryland. [FN14] A recent, widely-publicized attack in San Francisco has also brought the issue of aggressive dogs to the forefront of public attention. In January 2001, Diana Whipple was mauled to death by two Presa Canario dogs. The dogs were owned by a pair of lawyers. Evidence at the owners' murder trials showed that the dogs had tried to attack other people and animals in the past. Both defendants were convicted and served prison time. [FN15] A subsequent civil lawsuit brought by Ms. Whipple's mother was settled out-of-court. [FN16]

B. The Reaction of Some State Legislatures: "Breed-Specific Legislation"

  Highly-publicized Pit Bull attacks in the 1980s led to "knee-jerk reactions" by many communities. [FN17] Attacks led to editorials, which led to public outrage, which led to swift and spontaneous legislative action that was based on neither good science nor good law. "Breed-specific legislation" ("BSL") began to emerge in the 1980s and early 1990s. These laws targeted specific breeds for regulation or, in some cases, outright bans. BSL is on the rise in the United States. States and municipalities across the country have considered-and in some cases, enacted-breed-specific legislation designed to protect the public against dog bites. [FN18] Commonly, *8 these statutes and ordinances have banned, or placed restrictions on, Pit Bulls, Rottweilers, Doberman Pinschers, Chow Chows, German Shepherds, and Shar-Peis. [FN19]

  Ohio has aggressively targeted Pit Bulls for regulation. Ohio law declares any dog that "[b]elongs to a breed that is commonly known as a Pit Bull dog" [FN20] is automatically a "vicious dog." [FN21] "Vicious dogs" must be penned or tied up when on their owners' premises. [FN22] If off-premises, they must be tethered, caged, or muzzled. [FN23] Owners must obtain liability insurance to provide coverage in the event of a bite. [FN24]

  BSL has also occurred at the local municipal level. Denver, Colorado, passed an outright ban on the ownership, possession, keeping, control, maintenance, harboring, transportation, or sale of Pit Bulls. [FN25] A "Pit Bull" is defined as an American Pit Bull Terrier, American Staffordshire Terrier, Staffordshire Bull Terrier, or any dog displaying the majority of physical traits of one of those breeds. [FN26] This ordinance is in addition to Denver's "dangerous dog" ordinance which regulates "[a]ny dog with a known propensity or disposition to attack unprovoked, to cause injury or to otherwise endanger the safety of humans or other domestic animals." [FN27] "Dangerous dogs" must be confined while at home and must be leashed and muzzled while traveling. [FN28]

  Not all states have followed the BSL trend. Some legislatures have prohibited BSL enacted by municipalities. Florida enacted a statute that permits localities to regulate dogs "provided that no such regulation is specific to breed." [FN29] Some legislators have attempted, without success, to repeal this anti-BSL statute in response to several highly-publicized attacks. [FN30] Minnesota also has the following prohibition against BSL: "A *9 statutory or home rule charter city, or a county, may not adopt an ordinance regulating dangerous or potentially dangerous dogs based solely on the specific breed of the dog." [FN31]

  Court challenges to BSL have been largely unsuccessful. [FN32] Opponents of BSL have brought lawsuits claiming the legislation is unconstitutional because it violates due process (substantive and procedural), the Takings Clause, [FN33] equal protection, and the vagueness doctrine. [FN34] Plaintiffs have challenged BSL on due process grounds by arguing that there was no "rational relationship" to a legitimate legislative goal or purpose. [FN35] Courts have ruled that BSL is a rational response to a perceived problem of dog bites by certain breeds. [FN36] They have also rejected plaintiffs' arguments that the statutes and ordinances do not provide dog owners with sufficient notice and an opportunity to be heard, which are the requirements for procedural due process. [FN37] The Tijeras ordinance, for example, provides that a Pit Bull may be destroyed by the village only after a hearing to determine whether the dog is, in fact, a Pit Bull. [FN38] Plaintiffs have also contended that BSL amounts to a taking without just compensation. Courts have rejected this argument, noting that personal property is subject to *10 regulation under the police power of a state. [FN39] Challenges based on vagueness have argued that identifying a dog's breed is difficult. [FN40] Most courts have found BSL to be sufficiently specific to enable a reasonable dog owner to determine if his or her dog is covered by the particular statute. [FN41] Plaintiffs have also alleged that BSL violates equal protection by singling out Pit Bulls but not other breeds. [FN42] Courts have noted that Pit Bull ownership is not a "suspect classification" and therefore BSL need only have some reasonable basis to be constitutional. Courts have concluded that sufficient evidence exists to support a finding that Pit Bulls can be regulated by legislatures and municipalities. [FN43]

  One significant decision found BSL to be unconstitutional. In American Dog Owners Assoc., Inc. v. City of Lynn, [FN44] the Massachusetts Supreme Judicial Court upheld a trial court's finding that the City of Lynn's attempt to regulate Pit Bulls was unconstitutional. [FN45] The Court noted that it is particularly problematic to determine a dog's breed. The Court held, "[t]here is no scientific means, by blood, enzyme, or otherwise, to determine whether a dog belongs to a particular breed, regardless of whether 'breed' is used in a formal sense or not." [FN46] The Court upheld the trial court's finding that animal control officers had no real standards to identify Pit Bulls, in part because they had no training in breed identification. [FN47] The ordinance included a ban on mixed-breed dogs that contained "any mixture" of Pit Bull. [FN48] This provision was likewise found to be unconstitutional since it is scientifically "impossible to ascertain" whether a dog is part Pit Bull. [FN49] The ordinance was also unconstitutional because it tried to define "Pit Bull" as including any breed where "common understanding and usage" dictated that the dog was, in fact, a Pit Bull. [FN50] *11 The combination of these facts led the court to conclude that the statute was too vague to pass constitutional muster. [FN51]

C. The Reaction of Insurers: "Breed Discrimination"

  While some communities and states have responded to dog bites with breed-specific legislation designed to regulate or outlaw certain breeds, insurance companies have also reacted to the problem of dog bites in a breed-specific manner. Dubbed "breed discrimination" by dog owners, insurance companies have started making coverage and renewal decisions based on one's ownership of certain breeds of dog.

1. A Rise in Breed Discrimination

  During 2003 and 2004, the media brought breed discrimination to light. The CBS Evening News with Dan Rather aired a story in June of 2003 that featured a family that had difficulty obtaining insurance because they owned a Dalmatian. [FN52] The report stated, "[a]nimal lovers have a term for what the insurance company did. They call it 'breed discrimination'-arbitrarily punishing all dogs of certain breeds because some are vicious." [FN53] In the months that followed, several newspaper stories discussed the prevalence of breed discrimination and documented the effects this practice has had on families. [FN54] These news reports replicate the experience I had in *12 trying to get homeowners' insurance. Multiple insurers denied coverage because of the dogs I owned. I literally could not find a carrier in the Lubbock market willing to write a policy for me until I stumbled upon the Farm Bureau on the advice of one insurance broker who sympathized with my plight.

  The practice of breed discrimination produces absurd results. Consider the case of Chris and Norm Craanen of San Antonio, Texas. [FN55] They own a 12- year-old dog named Bukarus. He is a Rottweiler, a breed often targeted for discrimination by insurance companies. Yet, Bukarus does not pose much of a threat: he is deaf, partially blind, and has arthritis. [FN56] Despite his bite-free history, his owners lost their homeowners' insurance. [FN57]

  *13 Some of the most well-known insurers are engaging in breed discrimination. [FN58] Some insurers have outright bans on specific breeds, [FN59] while others take a more realistic and logical dog-by-dog approach. These decisions are predicated on insurers' assessment of relative risk. [FN60] The "usual suspects" for breed discrimination are Pit Bulls, Rottweilers, German Shepherds, Doberman Pinschers, Chow Chows, Wolf hybrids, and Presa Canarios. [FN61]

  The Humane Society of the United States has documented an increase in the number of people being denied insurance because they own certain breeds of dog. [FN62] As a result, the Society has started collecting data through the Internet, in the hopes of eventually convincing the insurance industry that there are alternatives to the current practice and that it must stop. [FN63] To achieve their goal, the HSUS and the American Society for Prevention of Cruelty to Animals have created a joint grassroots campaign designed to educate the insurance industry. [FN64]

2. The Insurance Industry's Defense of Breed Discrimination

  Homeowners' insurance protects a policyholder in the event of financial loss. Most policies include two provisions: property damage and liability. Property damage provisions protect the policyholder in the event of fire, lightning, wind, water or hail damage, theft, and vandalism. Liability provisions protect the policyholder in the event that a claim is made against a homeowner for negligence. Liability coverage typically pays for bodily injury, medical payments, and property damage that is *14 sustained because of the negligence of the property owner. [FN65] Absent breed discrimination, most homeowners' insurance policies would cover injuries due to dog bites on the premises between the amounts of $100,000 and $300,000. [FN66] In 1995, the average policyholder paid $418 in homeowners' insurance premiums. [FN67] By 2004, the average premium climbed to $608. [FN68]

  "Insurance is a business." [FN69] Insurers must make profits in order to continue in existence. [FN70] Companies survive by minimizing risk, which reduces the likelihood of claims. Some companies have decided that certain breeds of dog are simply "too much of a risk" to insure. [FN71] An industry representative claims that the issue of dog bites "is a major concern for insurers." [FN72]

  The industry defends its position, in part, on a series of studies from the Centers for Disease Control ("CDC"), which the industry claims as support for the proposition that certain breeds have a propensity to bite. [FN73] As I will demonstrate in Part II, however, the industry's reliance on the CDC studies is misplaced. Even the authors of the CDC studies have stated that breed discrimination is wrong and is not supported by scientific evidence. [FN74]

  The industry has also pointed to the large amount of money that has been paid out in recent years for dog bite claims. [FN75] The Insurance Information Institute ("III"), a trade group of the insurance industry, stated that in 2002, $345.5 million was paid out in dog bite liability claims, up *15 from $250 million in 1995. [FN76] The group argues that dog bite lawsuits are on the rise and juries are awarding larger claims. [FN77] They claim, therefore, the need to curtail their risk.

  The industry's cost statistics are misleading, however. The III states,  "[d]og bites now account for almost one quarter of all homeowner's insurance liability claims costing $345.5 million." [FN78] Some perspective is in order. For every $100 in premiums, insurers spend $77 paying claims. Of that $77, the overwhelming majority ($72, or 93.5%) is spent on paying property damage claims. Liability claims only amount to $5, or 6.5%, of total claims. [FN79] Even then, dog bites only constitute a percentage of that figure. Put into perspective, the money paid out in dog bite claims is negligible when compared to the overall amount of money paid out for other types of claims. Damage due to lightning, fire, and mold all individually account for more claims payouts than all liability claims combined. [FN80]

  The insurance industry has not been consistent in the reasons for its defense of breed discrimination. One report from III's website seems to defend breed-specific responses based on the aggregate claims paid [FN81] and stories of several high-profile and tragic bites. [FN82] However, in a statement to a newsletter of veterinary medicine, the III defended breed discrimination on the basis that certain breeds cause more damage when and if they do bite. [FN83] Ultimately, a spokesperson for the III conceded, "[t]he industry isn't *16 positioned to determine which dogs should be deemed vicious. . . . [W]e're certainly not dog experts or veterinarians." [FN84] This, however, has not stopped many insurers from engaging in breed discrimination.

3. Some Exceptions to the Rule?

  It appears that not all insurers have followed the breed discrimination trend. DVM reported that Nationwide Insurance changed its breed discrimination policy in October 2003. While Nationwide will now insure all dog owners, it will specifically exclude dog bites from its liability coverage. [FN85]

  State Farm's national representatives have repeatedly stated that the company does not practice breed discrimination. [FN86] However, when I searched for homeowners' insurance in 2003, a State Farm agent in Lubbock refused to even take my application because of the breeds I owned.

D. Other Instances of Breed Discrimination

  There are other examples where a person's ownership of a particular breed of dog can have negative consequences. Families seeking to adopt children can face roadblocks if they own dogs that belong to certain breeds. In Massachusetts, the Adoption and Foster Care Unit of the Department of Social Services will not place children in homes with certain breeds of dog. [FN87] The state relied upon data provided by the insurance industry when it made its decision to discriminate based on breed. [FN88] Some airlines also *17 practice breed discrimination by prohibiting some dogs from flying, even though they are stored in cargo and in a closed carrier. [FN89]

  Numerous scientific studies have attempted to identify the number of annual dog bites, the dogs most likely to bite, the people most likely to be bitten, and the circumstances under which bites are most likely to occur. Such studies have not reached a uniform consensus and have left us with more questions than answers. Even the studies that have attempted to report on breeds' proclivity to bite have cautioned that their research is incomplete and should not be used to justify breed discrimination by legislatures or insurers. [FN90]

A. CDC Statistics

  The U.S. Department of Health and Human Services, Centers for Disease Control and Prevention ("CDC"), commissioned a number of studies during the 1980s and 1990s to determine the scope and nature of the problem of dog bites in the United States.

1. Fatality Studies

  Four separate studies attempted to chronicle the number of fatal dog bites during the periods of 1979-1988, [FN91] 1989-1994, [FN92] 1995-1996, [FN93] and 1997-1998. [FN94] The studies were specifically limited to fatal dog attacks because fatality statistics are easier to track. [FN95] Non-fatal bites were excluded from the studies, although other scientists have attempted to use *18 emergency department reports and other sources to determine the number of non-fatal bites per year. [FN96]

  The authors combed three sets of sources in an attempt to determine the number of fatal dog bites per year. First, they searched NEXIS for news reports of dog bite-related fatalities. [FN97] Second, they used the National Center for Health Statistics' ("NCHS") single-cause mortality tapes ("SCMTs") to identify deaths where the underlying cause was listed as a dog bite. [FN98] Finally, the authors supplemented these reports with information collected by the Humane Society of the United States ("HSUS") to help identify the breed of dog involved in each incident. [FN99] From these three sources, the authors tried to piece together the number of people who died each year in the United States from dog bites.

  The authors concluded that dog bites caused approximately 7 deaths per year per 100 million people. [FN100] They discerned no identifiable trend that would indicate an increase in the incidence of fatal bites over the years of the studies. [FN101] During the first reporting period (1979-1988), approximately 70% of victims were under the age of 10. [FN102] Males, under the age of 29, were more likely than females to be victims. [FN103] These findings as to age and gender were consistent throughout the study periods.

  Many of the fatal bites of children involved horrific attacks on the very young. A 3-week-old girl was killed in her crib by the family's Chow Chow. [FN104] A 2-year-old boy in South Dakota wandered into a neighbor's yard where he was attacked and killed by two German Shepherd-wolf *19 hybrids. [FN105] The elderly were also victims of several fatal attacks. In March, 1996, two Rottweilers killed an 86-year-old Tennessee woman. One month before the assault, the dogs had attacked and injured the same woman. [FN106]

  In the twenty-year period of the CDC studies, the breed responsible for the most number of bites has changed. [FN107] From 1979-1980, Great Danes caused the most number of fatalities with three deaths for the time period. However, four breeds were tied with two deaths each: Pit Bulls, Rottweilers, Huskies, and Malamutes. [FN108] In 1981, Pit Bulls took over as the breed with the most number of fatal bites. [FN109] Pit Bulls remained in that position until 1993 when Rottweilers began causing approximately ten fatal bites per two-year reporting period. [FN110] The last available reporting period, 1997-1998, shows that Rottweilers caused ten fatal bites per two-year period, while Pit Bulls caused six, and Saint Bernards caused three. [FN111] During the 20-year study, 90 deaths were excluded because the breed was "unavailable." [FN112]

  The authors of the CDC studies acknowledged that the methods they used in their studies had a number of limitations. NEXIS, they pointed out, was not designed for scientific research. News reports would only be flagged if their text contained certain keywords. [FN113] Further, reliance on NEXIS assumes that newspapers accurately reported the breed of dog involved in a particular attack. [FN114] SCMTs have a one-to-two year lag time, which means that some fatalities may have been missed. [FN115] The authors believed that, on average, their methods only uncovered approximately 74% of dog bite related fatalities. [FN116]

  Even if one accepts the CDC statistics as definitive on the subject, they have a number of other limitations in answering the question of whether *20 certain breeds are more dangerous than others. First, the studies were limited to fatal dog attacks. [FN117] Secondly, the breed of the dog could not be accurately determined in every case. [FN118] Finally, the number of fatal attacks per year is so low that it is problematic to statistically extrapolate conclusions from the data. For example, in the first two years of the study (1979-1980), Great Danes accounted for the most number of fatal bites (three). [FN119] Four breeds, however, followed closely behind with two fatal bites each (Pit Bull, German Shepherd, Husky, and Malamute). [FN120] It would be statistically questionable to conclude that Great Danes were inherently more dangerous than the other breeds, based on a net difference of only one fatality.

2. Non-Fatality Studies

  The CDC fatality studies acknowledged that, while death rates for dog bites do not appear to have increased over time, [FN121] nonfatal bites were becoming more of a public health problem. [FN122] The CDC conducted a study on nonfatal dog bites in 2001. [FN123] The study used data from the National Electronic Injury Surveillance System-All Injury Program ("NEISS-AIP") to identify the number of nonfatal dog bites during the 2001 calendar year. NEISS-AIP collects data from initial visits to Emergency Departments ("EDs") across the country. [FN124] NEISS-AIP data is drawn from a nationally representative sample of NEISS hospitals. [FN125] The CDC analyzed every case where "dog bite" was listed as the external cause of injury. [FN126]

  In total, NEISS-AIP data revealed that hospital EDs treated 6,106 patients for dog bite-related injuries during 2001. [FN127] Since the NEISS-AIP *21 data did not include every hospital in the nation, the authors used this data to extrapolate to the general population. [FN128] They estimated that 368,245 people were treated for dog bite-related injuries in 2001. [FN129] The highest cohort of victims was children between the ages of five and nine. [FN130] Boys, under the age of 14, were more likely than girls to be seen in EDs for dog bite-related injuries. [FN131]

  The NEISS-AIP data included narratives for many of the attacks. One case involved a 4-year-old who was bitten by a dog guarding her puppies. [FN132] Another involved a 3-year-old girl who was bitten when she tried to take away a dog's food. [FN133] A 34-year-old man was bitten while trying to break up a dogfight. Some victims were bitten by their own dogs. A 27-year-old woman was bitten by her dog after he had been hit by a car and became disoriented. [FN134] A 75-year-old woman was attacked while trying to prevent her dog from biting an EMT who was attempting to put the woman in an ambulance. [FN135]

  The Morbidity & Mortality report describing the study does not document the number of attacks per breed. This is likely due to the fact that the ED reports did not specify the breed of dog. An attempt to determine the number of bites per breed would depend on victims accurately self-reporting the breed of the attacking dog. [FN136]

  The study had a number of limitations. First, the authors excluded fatal dog bites. Second, the study only examined cases where the victim sought treatment in an ED. Victims may have gone to other health care providers, such as private physicians or urgent-care centers. Third, 26% of reports were missing an injury diagnosis. Many cases had limited data on the circumstances of the attack or the identity of the dog involved. [FN137] Thus, the CDC's estimates may be both over inclusive ("just cause" bites may have been included) [FN138] and under inclusive (insofar as victims may have sought treatment at other facilities).

  *22 Another CDC study attempted to identify the incidence of dog bites in a particular locality: Denver, Colorado. [FN139] The authors examined reports from the Denver Municipal Animal Shelter in 1991. [FN140] There were a total of 991 bites during the study period. [FN141] However, only 178 were eligible for the study, [FN142] as the authors excluded several categories of bites: bites involving household members, attacks involving multiple dogs, attacks before 1991, dogs which had been owned for less than 6 months, cases in which the owner did not live in Denver County, attacks where the owner's phone number was not listed on the report, and cases in which the victim did not receive medical treatment. [FN143]

  The study created a control group of dogs to try to determine whether certain characteristics (such as breed) made a dog more likely to bite. [FN144] Using a multivariate statistical analysis, the study concluded that biting dogs were more likely than control dogs to be German Shepherds or Chow Chows, male, intact (not neutered), and reside in a house with one or more children. [FN145] Denver had (and still has) a ban on Pit Bulls, so it is not surprising that no cases involved that breed. [FN146]

  The authors acknowledged that their results had several problems. First, they were only able to speak to owners of approximately half of the biting dogs. They excluded cases in which the victim did not seek medical attention. In this respect, the authors believed that seeking medical attention was a "surrogate" for "real bites." [FN147] The authors did not verify the breeds of the dogs involved, but instead "identified predominant breed as whatever breed the owner considered the dog." [FN148] Because of the small number of bites per breed, the authors could not assess the statistical significance of breeds other than German Shepherd and Chow Chow. [FN149]

  *23 Another CDC study attempted to determine the frequency of dog bites by conducting a random telephone survey of households. [FN150] The authors used the Injury Control and Risk Survey ("ICARIS"), a random digit dialing telephone survey. [FN151] They asked each adult respondent whether he (or his children) had been bitten by a dog in the previous twelve months and whether the victim had sought medical attention. [FN152] Out of 5,328 completed interviews, 94 adults and 92 children reported being bitten in the previous twelve months. [FN153] Of these, 12 adults and 26 children sought medical care. [FN154] From this data, the authors extrapolated that 1.8% of the American population (4,494,083 people) had been bitten in the previous twelve months, and 0.3% had sought medical attention. [FN155] This shows that non-fatal bites are a public health problem that "is five orders of magnitude greater" than fatal dog bites. [FN156] The study concluded that several factors had no statistical significance on the likelihood of being bitten: census region, urbanicity, race/ethnic group, and household income. [FN157] The study did not attempt to correlate between the number of bites and the breed of dog. The authors acknowledged that the study relied on the self-reporting of data, which was not validated, and that they received a poor response rate (only 56% of people responded to the survey). [FN158]

B. Other Studies

  Other studies have attempted to document the total number of dog bites and the number of bites per breed.

  A study of ED visits for dog bite injuries [FN159] confirmed many of the conclusions of the previously-discussed CDC study on ED visits. [FN160] The study noted that a lack of a national reporting system for dog bite injuries makes gathering and analyzing data on the subject difficult. [FN161] The authors, *24 in reviewing the literature on the subject, found that previous studies concluded that between 0.3% and 1.1% of all ED visits are due to dog bite-related injuries. [FN162] To determine the true percentage, they collected data from the National Hospital Ambulatory Medical Care Survey ("NHAMCS"), a random surveying instrument that is used to calculate the number of ED visits per year. [FN163] They estimated that between 1992 and 1994, 333,687 annual visits were made to EDs seeking medical treatment for dog bite related injuries. [FN164] This amounted to 0.4% of all ED visits nationwide. [FN165] Looking at the monetary cost of dog bites, they found that the average cost for a dog bite related ED visit was $274, resulting in an annual cost of $102.4 million. [FN166] The study, however, did not address the question of whether certain breeds are particularly more dangerous than others. This is partly due to the unavailability of data through NHAMCS. Moreover, the study most likely undercounted the number of non-fatal dog bites because victims may have sought treatment from places other than EDs. [FN167]

  Other studies have attempted to examine the problem at a more localized level. A July 1991 study [FN168] found that dog bites were responsible for 0.3% of all ED visits at The Children's Hospital of Philadelphia. [FN169] Of those visits, 77% involved cases where the victim knew the biting dog. [FN170] The study found one statistically significant conclusion: more Pit Bull injuries were the result of unprovoked attacks as compared to such attacks by other breeds. [FN171] "Unfortunately, the absence of reliable dog breed-*25 specific population figures prevent[ed] the calculation of breed-specific injury rates." [FN172]

  An October 1997 study tried to determine the number of dog bites in Alleghany County, Pennsylvania (Pittsburgh) by using the "capture-recapture" method of statistical analysis. [FN173] The authors found that 790 dog bites were reported to the Alleghany County Health Department in 1993. [FN174] Using the capture-recapture method, along with log-linear modeling, the study concluded that the number of unreported dog bites was 1,388 (with a 95% confidence interval of between 1,010 and 1,925). [FN175] The authors cautioned, however, that the self-reporting sources are problematic in that "whether or not a case is reported depends largely on the severity of the event and the attitude, knowledge, or education level of the victim." [FN176] Accordingly, the authors suggested that the actual Pittsburgh dog bite incidence rate must be higher than that found in the study. [FN177]

  Another survey [FN178] in Pennsylvania polled children in order to determine an overall bite rate from the perspective of bite victims. [FN179] The survey, conducted in 1981, found that 46.1% of children reported that they have been bitten by a dog during their lifetime. [FN180] The study concluded that "being bitten by a dog is a rather common occurrence for children, especially those between the ages of seven and twelve years, and the event is greatly underestimated by official bite statistics." [FN181] Nevertheless, the authors did not attempt to catalog bites per breed. [FN182]

  Unfortunately, not all scientists have used statistically sound methods to draw conclusions about the relative dangerousness of breeds. Two physicians, Lee E. Pinckney and Leslie A. Kennedy, from the Department of Radiology at the University of Texas Southwestern Medical School and *26 Children's Medical Center sent letters to the editors of 245 major newspapers requesting copies of all stories about dog bite-related fatalities. [FN183] The number of fatalities reported by the responding newspapers between March 1966 and June 1980 totaled seventy-four. [FN184] Of the seventy-four fatalities, sixteen were caused by German Shepherds, nine by Huskies, eight by Saint Bernards, six each by Bull Terriers and Great Danes, and five by Malamutes. [FN185] The remaining dog bite fatalities were caused by a variety of breeds, including ten attacks by mixed breeds and five attacks by dogs of unknown breeds. [FN186] In addition to acquiring bite fatality statistics from newspapers, the authors used American Kennel Club ("AKC") registration data to compare the relative number of fatalities per breed. [FN187]

  The CDC authors criticized the Pinckney/Kennedy study as being "primarily anecdotal" rather than "systematic" in its approach. [FN188] Indeed, Pinckney and Kennedy conceded that their database was "incomplete" and "may not be entirely reliable." [FN189] Their data depended on newspaper reports, which may themselves be incomplete or inaccurate. Thus, the authors said their data required "cautious interpretation." [FN190] An example of such "cautious interpretation" is represented by the authors' observation that even though German Shepherds were involved in more fatalities than any other breed in the study, such large frequency could be reflective of the fact that German Shepherds had the highest AKC registration of any large breed. [FN191] Hence, the use of AKC data to draw comparisons between breeds is problematic, [FN192] as demonstrated by the high number of registrations for breeds such as German Shepherds, and low number of registrations for a popular breed, such as the Pit Bull. [FN193]

  William Winkler's study [FN194] in 1977 has also been criticized for its lack of scientific method. [FN195] His "study" involved compiling news reports from *27 eleven dog bite-related fatalities from January 1974 through December 1975. [FN196] From this data, he made various conclusions about the breeds responsible, finding that, "not unexpectedly," German Shepherds were the breed most often responsible for fatal dog attacks. [FN197] Because St. Bernards were responsible for two deaths during this 24- month period, he concluded, "[t]his relatively uncommon breed may be a greater hazard than others." [FN198]

  A common thread running through several studies is the attempt to extrapolate conclusions about breeds based on limited data. For example, an April 2000 epidemiological study in Philadelphia used reports from the Department of Health to conclude that between 1995 and 1997 there were approximately 5,390 bites. [FN199] The authors concluded that Pit Bulls, German Shepherds, and Rottweilers combined were responsible for 59% of bites each year. [FN200] The authors felt comfortable drawing this conclusion despite the fact that they could not determine the breed in 74% of cases. [FN201]

C. The Problem of the Unknown Origin of Aggressiveness

  Despite all of the research and studies on the subject, scientists and veterinarians cannot state with certainty or confidence why certain dogs are more aggressive than others. [FN202] It seems that a particular dog may be aggressive because of a variety of factors. [FN203] According to the American Veterinary Medical Association's multidisciplinary Task Force on Canine Aggression and Human-Canine Interactions, "a dog's tendency to bite depends on at least five interacting factors: heredity, early experience, later *28 socialization and training, health (medical and behavioral), and victim behavior." [FN204]

  While breed (as an inherited characteristic) is one component of predicting a dog's dangerousness, it is not the only factor. [FN205] There is no way to scientifically determine whether a dog is likely to bite in the future, anymore than psychologists can predict whether certain people will commit crimes of violence. The exception to this rule is the axiom that the best predictor of future behavior is past behavior. For this reason, many veterinary and scientific groups support "dangerous dog laws" which target individual dogs that have demonstrated a propensity to bite or attack innocent victims. [FN206] The problem with BSL and breed discrimination is that legislatures and insurers have attempted to prophylactically determine which breeds are most likely to bite without any evidence of individual dangerousness.

D. The Problem of Numerators and Denominators in Dog Bite Statistics

  To date, no scientific study has been able to resolve what I term to be the problem of "numerators and denominators." A person wishing to determine whether certain breeds are more likely to bite than others must first determine the number of bites per breed (the numerator) and then compare that number to the total number of dogs of that breed in the general population (the denominator). This can be expressed as a ratio:

                                 Number of Bites by Breed

Relative Dangerousness Ratio  =  -------------------------

                                 Total Population of Breed

  This ratio ("RDR") allows for a comparison between breeds. The higher the RDR, the greater proclivity a particular breed has to bite. It allows for a comparison of "oranges to oranges" and "apples to apples." Otherwise, it is likely that highly popular breeds will appear to be more *29 dangerous, when in fact the number of bites is reflective of the overall population of the particular breed.

  A study which tried to extrapolate breed data from the previously-discussed CDC studies agreed that the proper method for determining a breed's dangerousness was the use of a comparative ratio:

    Ideally, breed-specific bite rates would be calculated to compare breed and quantify the relative dangerousness of each breed. For example, 10 fatal attacks by Breed X relative to a population of 10,000 X's (1/1,000) imply a greater risk than 100 attacks by Breed Y relative to a population of 1,000,000 Y's (0.1/1,000). Without consideration of the population sizes, Breed Y would be perceived to be the more dangerous breed on the basis of the number of fatalities. [FN207]

  Using the RDR normalizes the effect of a breed's popularity, or lack thereof. Dogs of popular breeds are going to bite more often simply because there are more of them. [FN208] A January 1997 article warned that, as Dalmatians become more popular, people should expect to see more bites from that breed. [FN209] This is not to say that Dalmatians are inherently more dangerous than other breeds. Rather, an increase in their population should also result in a proportional increase in bites from that breed. [FN210] Similarly, the Pinckney/Kennedy study [FN211] cautioned that, despite the fact that German Shepherds accounted for the most number of deaths, their finding must be read in conjunction with the popularity of the breed, as evidenced by AKC registrations of the same time period. [FN212]

  The problem of numerators and denominators is that it is difficult-if not impossible-to accurately determine the number of bites per breed and *30 the number of dogs in a particular breed. Without an accurate count for either the numerator or denominator, one runs the risk of stigmatizing an entire breed as "overly dangerous" based on the breed's absolute number of bites, instead of examining the breed's number of bites relative to its overall population.

1. The Numerator Problem

  The principal problem in determining the total number of bites by a particular breed is that there is no national reporting system for dog bites. [FN213] The CDC studies [FN214] demonstrate that while fatal dog bites are easier to track than non-fatal bites, even the methodology used to uncover fatalities misses approximately 26% of cases. [FN215] Further, news accounts-on which the CDC relied, in part, to determine the number of fatal dog bites and the breeds involved-may be biased towards reporting attacks by certain breeds. [FN216]

  The numerator may also be biased against dogs that cause more damage, while ignoring breeds that bite more often but do not cause victims to seek emergency treatment. [FN217] If a dog bite does not cause serious injury, it is not likely that the victim would seek medical treatment. [FN218] This then skews the results of studies that use emergency department visits to track the incidence of dog bites. [FN219] "The problem with self-reporting sources is that whether or not a case is reported depends largely on the severity of the event and the attitude, knowledge, or education level of the victim." [FN220] Studies which have used random sampling [FN221] are equally problematic because they, too, depend on accurate self-reporting of their sample groups. The low response rates of these studies also lead to questions about the accuracy of the results that are extrapolated to the general population. [FN222]

*31 2. The Denominator Problem

  No one knows how many dogs are present in the United States at any one time. This should not be surprising, as even the Constitutionally mandated [FN223] decennial census of human beings is known to undercount people. [FN224]

  Determining the true number, or even an accurate estimate, of dogs can be problematic. While many dogs are kept as household pets, [FN225] others are used as service animals or guard dogs; kept in animal shelters or animal stores; or simply allowed to wander the streets as strays. The dog population is constantly changing and moving, which makes obtaining an accurate count difficult and expensive.

  Even if it was possible to determine how many dogs exist in the country at any one time, the problem then becomes how to determine how many of those dogs belong to each breed. Determining the breed of one dog is difficult enough. [FN226] To take a census of all dogs and identify their breeds would be an impossible task.

  Some scientists have suggested using AKC or municipal registration data to determine the number of dogs in a particular breed in a particular community. [FN227] However, one study concluded that city registrations account for only 29.1% of all dogs. [FN228] Further, owners of breeds considered "dangerous" may be reluctant to register their animals. [FN229] This may be particularly true of dogs used for illicit purposes, such as those owned by drug dealers, dog fighters, and gang members. [FN230]

  *32 American Kennel Club ("AKC") registration data is also problematic because the AKC only registers pure-bred dogs [FN231] and depends on owners taking the initiative to register their dogs. [FN232] Mixed breeds, for which there are numerous combinations, are not eligible for registration. [FN233] Pit Bulls are often registered with organizations other than the AKC. If owners do register them, they register with the United Kennel Club or the American Dog Breeders Association. [FN234] If a breed is undercounted in the denominator of the ratio, it will make a breed appear more dangerous than it actually is. [FN235]

E. The Problem of Breeds

  Breed is a human construct that is used to conveniently group dogs based on similar physical characteristics. [FN236] There is no scientific test to determine a dog's breed. [FN237] The only way to determine a dog's breed is to examine its heredity. This task is made possible but expensive and time-consuming, [FN238] if a dog is registered with the AKC. [FN239]

  As examples of the problem of defining and identifying breed, consider the case of Huskies and Pit Bulls. "Husky" refers to a class of dogs, not any one particular breed. Siberian Huskies, Alaskan Malamutes, and Samoyeds are all considered to belong to the "Husky" family, yet they are all different breeds. [FN240] Similarly, there is no AKC-standard breed called *33 "Pit Bull." "Pit Bull" is a collective classification of the American Staffordshire Terrier, Staffordshire Pit Bull Terrier, and Bull Terrier. [FN241]

  Scientists have not been able to determine if victims of dog bites can accurately report the breeds of dogs that attacked them. Many scientists, particularly the CDC authors, have stated that misidentification is a likely problem, especially under the stress of a dog attack. [FN242] Part of the problem may be that as a particular breed gets a reputation for dangerousness, some victims jump to the conclusion that they were bitten by a dog of that breed. [FN243]

  Even under ordinary, low-stress conditions, many people have difficulty identifying a dog's breed. "For the average person anything with prick ears and blue eyes automatically becomes a 'husky' . . . Any smooth coated brown dog, medium sized, and muscular becomes a 'pit bull' . . . Any tall dog becomes a Great Dane, fuzzy or hairy and it's a Chow Chow. If it's black and tan and heavy it's a Rottweiler, etc." [FN244] One survey of bite reports found that medium-sized black and tan animals were likely to be recorded as German Shepherds. Stocky, short-haired dogs were listed as Pit Bulls. Media reports of Pit Bull attacks are often accompanied by pictures of Boxers or Pugs instead of American Staffordshire Terriers. [FN245] One entertaining website called "Find the Pit Bull" displays 21 pictures of pure-bred dogs and challenges the user to identify the Pit Bull amongst them. [FN246]

  Even veterinarians and other experts have difficulty determining whether a particular dog belongs to a particular breed. [FN247] This was a central concern of the Massachusetts Supreme Judicial Court in American Dog Owners Ass'n v. City of Lynn. [FN248] The Court declared the city of Lynn's Pit Bull ordinance to be unconstitutional in part because the animal control *34 officers designated to enforce the ordinance used conflicting and subjective standards to determine and identify breed. [FN249]

  The problem of mixed breed complicates the issue even further. In determining a relative dangerousness ratio, it is unclear how to count mixed breeds. [FN250] Should they be counted once per breed? Not at all? Create a new category for each possible combination of breeds? Aside from how to use the raw data on attacks by mixed breed, there is the additional problem of misidentification by lay people. [FN251] Victims sometimes inadvertently report mixed breed dogs as pure-breds [FN252] due to the heat of the moment and their lack of training in identifying subtle breed characteristics.

  There is good reason to believe that the raw data being used to calculate relative dangerousness ratios is incomplete and inaccurate. If the data being inputted into the calculation is flawed, the results (claiming to show some breeds are more dangerous than others) are equally flawed. [FN253]

F. The Problem of "Just Cause" Bites

  Even if an accurate count could be obtained of the number of bites per breed, there is the additional problem of how to handle "just cause" bites in the resulting statistics. If the purpose is to determine which breeds are inherently more dangerous, just by virtue of the breeds themselves, then the statistics should exclude bites that were justified by the dog. If a Rottweiler bites an intruder who is attacking the homeowner, we would expect the Rottweiler to be praised for defending its owner. This is not the type of bite that we should be trying to prevent. It is also not the type of bite that is likely to lead to an insurance claim. Similarly, if a dog is being physically tormented by a neighborhood child who is poking it in the eye, we would not deny that the dog has an inherent right to defend itself by growling, snarling, barking, or biting back. [FN254] These are "just cause" bites, bites in which the dog has a legitimate reason to defend itself or its owners.

  *35 It is possible that the statistics are being skewed because property owners who wish to purchase "guard dogs" may be self-selecting certain breeds based on the popular notions of relative dangerousness. Guard dogs are trained to protect property by scaring away would-be intruders and, if necessary, to bite an actual trespasser. Owners who desire to have guard dogs may rationalize the purchase of one breed over another based on the degree to which they subjectively believe that the dog will be "mean" or "scary." This creates a self-fulfilling prophecy. The "scarier" a breed is considered by a community, the more likely a dog of that breed will be purchasedfor protection, used for protection, and actually bite an intruder. This will skew the statistics in a way that purports to show that the particular breed is, in fact, inherently more dangerous.

  Despite these concerns, it appears that the studies to date have not excluded this category of bites from their datasets. [FN255] This is a fatal flaw in the statistics, for it confuses the issue between inherent dangerousness (due to breed) and legitimate animal behavior.

G. The Problem of Breed Switching by Dog Fighters, Gang Members, Drug Dealers, and Other Bad Owners

  Assume for the moment that an accurate relative dangerousness ratio could be determined for each breed and that it could be scientifically determined that certain breeds are inherently more dangerous than others. What about the owners? Does this not excuse them from the responsibility to properly train and care for their pets?

  The reality is that there is a wide spectrum of responsible pet ownership. For some people, occasionally providing food and water for a dog is considered sufficient. On the opposite end of the spectrum, some people spend thousands of dollars on luxuries such as pet spas, advanced dog agility classes, and elaborate beds. Somewhere in the middle of the spectrum are people who actively ensure that their pets have food, water, and shelter; get exercise; are well-trained; and receive adequate veterinary care. [FN256]

  Unfortunately, a small percentage of pet owners breed and use their pets for illicit purposes. They intentionally seek out vicious dogs that will attack and maim humans and other animals. [FN257] Dog-fighting enthusiasts, gang members, and drug dealers will purposely select, breed, and train *36 dogs to be vicious. The purpose may be to intimidate rivals (in the case of gangs and drug dealers), to defend illegal drugs (in the case of drug dealers), or to make money (in the case of promoters of dog fights). [FN258] For some, having a vicious dog is simply a status symbol. [FN259] In order to make dogs into vicious weapons, they use "revolting and painful techniques to bring the animals to the verge of bloodlust." [FN260] Drug dealers in Philadelphia during the 1980s had Pit Bulls named "Murder, Hitler, and Scarface." [FN261] They wore collars that concealed crack, cocaine, and money. [FN262] In Chicago, gang members "brandish[ed] their fierce pit bulls just as they would a switchblade or a gun." [FN263]

  Current statistics do not take into consideration the degree to which the source of a dog's aggressiveness is the torturous upbringing described above, as opposed to the dog's breed. [FN264] In those situations, the problem is clearly with the dog owner-not the dog itself or its breed. These problem owners are dangerous with any breed of dog. [FN265]

  One solution would be for insurers to write policies that exclude injuries related to dog fighting. This would limit the claims paid out for these high-risk animals, yet it would leave potential plaintiffs without an adequate source of compensation. This result might be a socially acceptable solution because of the unclean hands of the "victims." If dog-fighting exclusions are incorporated into standard homeowners' insurance contracts, the language should be narrowly written to exclude only those bad faith actors who, as a matter of social policy, should not be rewarded or compensated for injuries attendant to an illegal activity. The key would be to write language that would still protect innocent passers-by.

  One of the arguments against BSL is that once a breed becomes banned, problem owners will simply switch to another breed. [FN266] In the 1930s, Pit Bulls were far from considered a "vicious breed." In fact, a Pit Bull named "Pete" starred in the Our Gang films of the time. [FN267] Fifty years ago, the Doberman was considered the most vicious dog. [FN268] During the *37 1980s, the focus turned to Pit Bulls. [FN269] In short, today's public target may be tomorrow's favorite pet, and vice versa.

H. Do the Insurance Companies Have Better Data?

  It is quite possible that one or more insurance companies have their own proprietary data purporting to show that one breed or another is disproportionately responsible for bites. I am skeptical that their data would be any better than the CDC's. The problems associated with the CDC and non-CDC studies are inherent to the problem of trying to determine the number of bites per breed and the number of dogs per breed.

  Breed discrimination by insurance companies and breed-specific legislation by state and local governments have attracted national attention and outrage by veterinarians, animal groups, and dog owners.

  The American Veterinary Medical Association's Task Force on Canine Aggression concluded that BSL and other breed-specific actions are "inappropriate and ineffective." [FN270] The Task Force consisted of a diverse coalition of veterinarians, academics, physicians, insurers, representatives from animal rights advocates, CDC scientists, and lawyers. [FN271] The Task Force agreed that to properly determine the relative dangerousness of breeds, one must first determine the number of bites per breed and the total population of each breed. As noted above, [FN272] the accurate calculation of both numbers is an immense challenge. [FN273]

  The Task Force rejected the notion that a dog's breed is the sole determinant of dangerousness. "[A] dog's tendency to bite depends on at least 5 interacting factors: heredity, early experience, later socialization and training, health (medical and behavioral), and victim behavior." [FN274] They also pointed to the problems of mixed breeds, misidentification of breeds, and shifting popularity of breeds. [FN275] The Task Force also expressed concern about making decisions based solely on breed, since there is a lack *38 of scientific means to identify breed. [FN276] The Task Force recommended, instead, that local governments focus on individual dogs and dog owners. [FN277]

  The very scientists who have authored studies trying to determine a link between breed and aggressiveness oppose breed discrimination and BSL. In many of the CDC studies, the scientists cautioned against using their incomplete data on attacks to make knee-jerk legislative or policy decisions based solely on breed. [FN278] They pointed to the lack of reliable data on bites per breed (the "numerator problem") and the absence of a reliable count of dogs per breed (the "denominator problem"). [FN279]

  Animal groups have also opposed BSL and breed discrimination. The American Kennel Club ("AKC") has taken a strong stance against breed discrimination by insurance companies:

    The American Kennel Club believes that insurance companies should determine coverage of a dog-owning household based on the dog's deeds, not the dog's breeds. If a dog is a well-behaved member of the household and the community, there is no reason to deny or cancel coverage. In fact, insurance companies should consider a dog an asset, a natural alarm system whose bark may deter intruders and prevent potential theft. [FN280]

The AKC also issued this statement concerning BSL:

    The American Kennel Club (AKC) strongly supports dangerous dog control. Dog control legislation must be reasonable, non-discriminatory and enforceable as detailed in the AKC Position Statement.

    *39 To provide communities with the most effective dangerous dog control possible, laws must not be breed specific. Instead of holding all dog owners accountable for their behavior, breed specific laws place restrictions only on the owners of certain breeds of dogs. If specific breeds are banned, owners of these breeds intent on using their dogs for malicious purposes, such as dog fighting or criminal activities, will simply change to another breed of dog and continue to jeopardize public safety. [FN281]

  In response to a perceived rise in breed discrimination, the Humane Society of the United States ("HSUS") and the American Society for the Prevention of Cruelty to Animals ("ASPCA") developed a grassroots campaign to educate the insurance industry. [FN282] Both groups oppose breed discrimination. [FN283] Other groups that have spoken out against breed discrimination include the American Medical Veterinary Association, the American Dog Owners Association, the Westminster Kennel Club, and the American Humane Society. [FN284]