Motivation

Pit bulls have a historically poor reputation as an aggressive breed. They’re most commonly used in dog fighting, and are often not permitted as pets in many leasing communities. Because of their poor reputation, thousands of pitbulls are euthanized in kill shelters each year. Supporters of the breed have championed these dogs as sweet and loving in nature, and in recent years have been pushing to change the negative stigma that surrounds them. Below are some resources that dive into these topics, from both organizations and individuals:

The purpose of this project was to use quantitative data to assess whether pitbulls are more aggressive than other breeds. By using reporting of dog bites as a surrogate measure of aggression, can we see whether pit bulls are more aggressive than other breeds, and whether this has changed over time?

A screencast of our project can be found here. A more fun version of this screencast can be found here - with music! Please enjoy both.

Data

We used data publicly available on NYC Open Data. More detailed information on the datasets used can be found in the Data tab. A Google Drive folder of the datasets can be found here.

Research questions

Main results

Discussion

Quantifying dog aggression with individual-reported dog bite data in NYC provides insights into trends across breeds, locations, and over time. It was interesting to note that pit bulls account for such a high percentage of reported dog bites, especially since they make up such a small proportion of all registered dogs. It begs the question: why? Is this because pit bulls bite people more often, or is it a matter of people being more inclined to report bites by pit bulls versus other breeds? Unfortunately, this is something we can’t answer.

Still, we found that pit bull bites increased over time at a much lower rate - almost not at all- compared to bites from other breeds. If this is a facet of reporting, then it may be that efforts to change the stigma surrounding pit bulls is working, and people are reporting such bites less often.

The limitations of this dataset make inference difficult. The opt-in data collection makes for potentially biased data. For example, dog sex was largely unknown, calling into question the meaningfulness of our conclusions regarding the associations between dog bites and gender. Despite this, we think our findings are interesting, and at the very least prompt further investigation regarding pit bull behavior and stigma.