Quite a few of the assumptions that the Puppycide Database Project team shared about police use of force toward animals have been shattered during the course of our research. One of the original catalysts of this project was an interest in criminal justice scholarship concerning the militarization of United States domestic policing over the last 50 years. Many of the earliest indications that police killing of pets was more frequent than the public understood were reports and videos of SWAT team raids being leaked; in many such incidents, killing any dogs at the scene appeared to be part of common practice, regardless of breed or other options for containment. As such, we fully anticipated that a majority of the puppycides that we found would be caused by the use paramilitary police tactics during the serving of drug related warrants.

The trends that we have uncovered to date reveal a much different reality: one in which paramilitary-style tactics during warrant seizures regularly result in the deaths of animals, but where the vast majority of "puppycide" occurs either in the middle of the street in a residential area, or when a police officer enters a home (or surrounding property, such as a fenced back yard) without a warrant or permission from the owner. Many such proprerty invasions rely on exemptions to the Fourth Amendment that are very recent and highly controversial; exemptions that police use to search entire neighborhoods without a warrant under the belief that a single criminal fled into that neighborhood. A great deal of these property invasions that result in the deaths of dogs do not appear to meet any current constitutional exemption, and appear to be flatly illegal. In the small number of these cases that have made it all the way to a civil trial, the illegality of many of these unwarranted and unannounced searches has become apparent, and has lead to large monetary awards to the families of the victims.


The statistics used in this article are based on a survery of 914 incidents involving the use of deadly force toward animals by police between the years 1990 and 2014; a total of 1322 animals were either killed or seriously injured during these 914 incidents. We say "incidents" because police relied on several different means to injure, maim and kill animals, including but not necessarily limited to - stabbing animals with knives, bludgeoning animals with a blunt instrument (brick, etc), running over animals with a squad car, kicking animals in the skull while wearing "combat" boots, asphyxiating animals with wire, as well as the more widely known instruments such as baton, electrocution via taser, and chemical weapons to cause severe pain such as CS gas and "mace".


During the course of these 914 incidents, in only 214 cases or 23.4% of total cases did police have a warrant or explicity permission to enter private property. All of the cases involving paramilitary-style raids are included in this 23.4%, but these also include several incidents in which a person lost control of their animal and requested assistance from the police, who in turn killed the animal. This also includes a handful of cases where zoos or nature preserves contacted the police for assistance because an animal escaped their cage or where someone entered an an animal cage, and the police in turn killed the animal. These latter groups are important; cases where police had permission to enter the premises represented 35.2% of all cases where animals caused injuries resulting in hospitalization or death to a human being. Conversely, situations where animals caused injuries resulting in hospitalization or death to a human being represented 17.8% of all cases where police had a warrant or permission to enter private property - a much higher rate of animal-caused injury than the total survey group, for which the rate of hospitalization or death was 11.8%.

Incidents in which the use of force incident took place on public property was the largest such group examined, at a rate of 36.2% of total incidents. By "public property" we rely on the standard meaning, of property either owned explicitly by the state, like a government building, or owned through some sort of "commons" theory of ownership, like a public park. In addition to this standard meaning, we add two provisos for the sake of brevity. Property that is privately held through a homeowners association we have deemed "public" for the purposes of our examination. Additionally, property that is held through a commons but assigned some private maintenance contract or right of exclusive use is here still considered "public". This allows us, despite the theoretical inaccuracy of such a claim, to hold that all uses of force that occur on a street, a sidewalk and most types of parking lot are "public". This saves our researchers the additional responsibility of reviewing a multitude of contracts for such things, which presents a substantial burden while only serving to needlessly complicate the results.

The next group is uses of force in which it is disputed or otherwise not currently known is police had a warrant of explicity permission to enter private property during the incident in which deadly force was used against an animal. This troubling group accounts for 12.3% of total cases. Some examples that lead an individual incident into inclusion in this group are the following: cases where the family of the victim and the police officer contradictory accounts of the shooting or where it occurred, cases where police have refused to state publicly where the deadly force was used and requests for internal police documentation have not yet been processed and cases in which police have offered multiple, contradictory versions of events or in which police statements are otherwise rendered uncredible. It is worthwhile to note here that we do not include in this group cases in which a warrant has been contested in court or even cases in which a warrant has been deemed illegal by a court. While such information is absolutely worth reviewing, it is simply beyond the bounds of our current examination. Thus, if any actual warrant exists, even a warrant later deemed illegal, we include such an event as a warranted/permitted search.

The last grouping involves cases in which police invaded private property, used deadly force against an animal found within that private property, and did so without any form of warrant of permission from the owners of that property. This group accounts for a stunning 28.4% of total incidents included in the survey. This group is significantly larger than the first group we looked at, which included not just paramilitary-style "SWAT" raids, but also cases in which police were invited into the home at the behest of the owner - that combined group accounted for only 23.4% of the total. It is vital to note here that this group includes searches of private property for which there exists a legal exemption to the Fourth Amendment and situations where police conduct a search that is illegal or whose legality is disputed. So for example, cases which rely on a so-called hot pursuit exemption allowing police to chase a suspected felon through private property are included here, but so are cases where police entered private property on the suspicion that drugs were inside the home. In many cases included in this group, police have flatly stated that police entered a private home without a warrant and offered no further justification whatsoever.

Let's consider the "disputed" group in the context of the "warranted" and "unwarranted" groups for a moment - at over 12%, wouldn't at least a portion of the "disputed" group be applied to groups with warrants, once additional information becomes available? While it is possible that some of the disputed group may in fact turn out to have had warrants all along, and police for whatever reason chose to keep that information secret, it is unlikely to account for enough of them to make the "warranted" group larger or even in parity to the "unwarranted" group. This is because, in the majority of disputed cases, the dispute resides in whether the shooting occurred on public or private property (with police claiming the shooting occurred on public property and the victim family claiming the shooting occurred on private property). If additional information becomes available confirming the police's point of view, it would merely add such cases to the "public property" group, leaving the current standing between the "warranted" and "unwarranted" groups unaffected.

This consideration brings us back to the unexpected result that began this essay: SWAT teams are responsible for only a minority of police use of deadly force against animals. The majority of "puppycide" is being carried out by patrolmen who either kill the animal on a public street or sidewalk, or who enter private property without a warrant and kill an animal they encounter during that unwarranted search. Together, these two scenarios account for at least 64.7% of police use of deadly force toward animals. What is the driving force behind these deadly encounters? What begins the situation that leads to police either shooting an animal on a street or sidewalk, or invading private property without a warrant and using deadly force against the animals they encounter?


In the case of police shooting a dog on public property, one narrative overwhelming accounts for the majority of uses of deadly force. The narrative begins with a member of the public calling 911 upon seeing a dog they believe to either be a stray or to have run away from home. Rarely is it the case that such dogs are engaging in violent behavior. While many such 911 calls involved the caller describing the dog as "menacing" or "threatening", such descriptions appear to be unrelated to the dogs actual behavior in these cases; callers describing a dog as "menacing" are in reality communicating to police that the caller is afraid of dogs, the caller has seen a dog, and the dog is not acting violently toward the caller. Based upon the examination of the outcomes of these calls, the caller represents a much greater threat to the dog than the dog does to the caller (of the 914 cases examined, none of those who called 911 died, while 79.7% of the cases resulted in the death of at least one dog). Police arrive on the scene, and proceed to shoot the dog.

Occasionally, police responding to 911 calls from those reporting a loose/stray dog do not immediately find the dog. When this happens, we have recorded many incidents in which police proceed to engage in searches of private property without a warrant, even in cases where the dog has not attacked or injured anyone. These searches tend to take the form of the police looking into back yards and alongside houses for the dog. In many states, such searches are considered illegal when the dog has not acted violently. Some have argued that any stray dog presents a violent threat to the community, whether or not that dog has actually attacked anyone. While community standards and beliefs on that matter may differ, the annual actual number of deaths caused by dogs are extraordinarily low, with 34 such deaths in 2011. That number is much lower than, for instance, the annual number of deaths caused by police, which by even the most conservative estimates provided by the FBI, which are self-reported exclude a large amount of officer-involved shootings, exceed 400.

These "on-the-spot" searches account for only part of the total number of searches performed without a warrant that resulted in police using deadly force against an animal. The second scenario, which we found to account for a large number of incidents, involves situations in which police claim to be in "hot pursuit" of a suspected felon. During such pursuits, police proceed to perform searches of a large number of homes throughout an entire neighborhood in order to apprehend the suspected felon. Occassionally such pursuits involve dozens of police, on foot, in squad cars and even in helicopters and other types of aircraft. When these pursuits are recorded by the Puppycide Database Project, it is because a police officer involved in the pursuit runs into the yard or home of a family that, in almost every case, is not suspected of any involvement in the crime. It is important to stress that point - that in very nearly every "puppycide" case involving a hot pursuit, police enter the private property of a family that the police themselves claim is in no way involved with any criminal behavior. Once the pursuing police officer enters the home or yard, an animal is encountered; sometimes that animal is secured by a chain or leash, sometimes the animal is simply secured within its own home or fence (which the police have now entered). Police than proceed to use deadly force against an animal.

It is of interest to note that in the subset of cases where police did not have a warrant, (those cases involving "hot pursuit" are included in this subset), the odds of the animal involved causing death or serious injury to a human being is significantly less than the overall survey group: for the overall survey group, animals caused death or serious injury in 11.8% of cases. However, animals caused death or serious injury in only 4.2% of the cases where police entered private property. As mentioned earlier, when police had a warrant or were given permission to enter by the property owner, animals caused death or serious injury in 17.8% of cases.

This result has substantial statistical significance, for which there are a number of possible explanations. One possibility is that police shootings of animals during "hot-pursuit" of a felon, or in situations where police must essentially "hunt down" the dog onto private property are cases in which the police themselves are resposible for the escalation of violence; in these situations, police are using violent force against a dog that has not caused serious injury in a full 96% of reviewed cases.

Another way of reviewing such findings is that police are themselves responsible for the escalation of violence in a majority of cases, whether a warrant exists or not. To explain: the "warrant and permission" group includes cases in which paramilitary tactics are used. By their very nature, such tactics involve the escalation of force by police. However, when taking this view we must account for cases in which police are called and invited into a home only after a dog has caused serious injury or death. We must also account for a smaller group of cases in which paramilitary tactics were justified by a previous act of violence not involving a dog, but which results in the death of a dog. For example, we recorded a number of events in which a police SWAT team was deployed after police were informed that someone had shot or stabbed someone and then barricaded themselves in a home; the SWAT team then proceeded to shoot a dog who had nothing to do with the situation, but who was merely present at the scene. A larger group of SWAT team deployments recorded in our survey were for individuals who were suspected of non-violent crimes, but who police believed had been "barricaded" into a home. While the police escalation of force is easily debatable in all of these situations, in at least some of them a SWAT team may have been otherwise appropriate, even if the SWAT team's subsequent killing of an animal was not.

There is at least one other way of reviewing these results, as well. In cases where the rate of human injury and death is low, but the rate of lethal force used against animals is high, we can conclude that police are doing a better job than in cases where the rate of human injury and death is high and the rate of lethal force used against animals is high. After all - are police supposed to wait until someone is seriously injured or dead until they act? Surely police officers are bound by a responsibility to protect the public from harm. The fact that searches without warrants resulting in lethal force against animals have a low correlation with injuries caused by dogs could simply mean that police are more alert during such circumstances, and more likely to act before the animal is able to get ahold of the police or a bystander with tooth and claw. This is certainly the opinion favored by police themselves.

While it is reassuring to see results of lower injury rates as advantageous, it is also a view that is deeply problematic. For us to see smaller rates of injury as a success, we must take for granted that a higher rate of injury would occur, absent for the use of deadly force. Only under those circumstances could we see police use of force as "protection" or "self-defense". What if, under the same circumstances, we were to remove the use of force entirely, and the rate of injury were to stay the same? What if, absent the use of deadly force, the rate of injury were actually lower? If either of those situations were to be true, than we would be forced to cease considering the use of deadly force as "protection" or "self-defense" and assign it a much uglier label entirely (the exact label is left as an exercise for the reader). In order for us to determine the legitimacy of this assumption, we would require a very specific type of control group. In a prior research project, the Puppycide Database Project was able to determine the rate of death and hospitalization caused by animal bites and scratches for the entire United States over a long period of time. While this might seem like a sufficient control group, in actuality we would need to review the rate of injury caused by animals for police searches conducted without a warrant. The reader can easily imagine some of the difficulties in obtaining such a survey, not least of which being that police themselves have made such data extremely hard to come by - internal police documents are handled as if they were classified, despite the fact they are legally public record and should be readily available to any US citizen. It is this lack of transparency which has necessitated the Puppycide Database Project in the first place. It is somewhat ironic, then, that it is this same lack of transparency that makes it nearly impossible to confirm the view held by most police departments on this topic with anything resembling scientific rigor.

Continuing Publication

This blog post represents the first public review of the research gathered by the Puppycide Database Project. As we continue to add to our data and our analysis of that data, we will continue to provide "Early access" to our results through blog posts like this one.

In the subsequent posts, to be published within the next week or so, we plan to release short articles expanding some of the discussion here. Like today's post, many of these will discuss areas which resulted in a clear divergence from the results we expected from our research. Here are a few such topics:

  • How often are humans killed or hospitalized when police attempt to use deadly force against an animal, and miss?
  • How many children are caught in the "line of fire" when police shoot an animal with a gun?
  • What are some of the current issues with Puppycide Database Project research methodology? How do we plan to address such issues, and how do they impact our current findings?

As always, feedback from all Puppycide Databae users, readers and Volunteers is highly encouraged. You are also welcome to use Puppycide Datahase Project research in your own publications, so long as you provide a citation (a link is also appreciated, but not required). Best of all, anyone can review our analysis simply by requesting the current Puppycide Database.