Gun Policy Effectiveness Predictor - UC Berkeley MIDS
Explore State Statistics
We can begin an analysis by evaluating statistics in each state. The visual below illustrates the "crude rate" of gun deaths per 100,000 people annually in each state. Using your cursor, you can highlight the states that you are interested in looking at by drawing a box around it (and clicking "keep only"), or alternatively, use the selection box to determine which states you'd like to evaluate in more detail. If you hover over the upper left corner of the map, there are additional links for a navigation menu that provides more options for zooming, scrolling, or searching.
The state with the highest number of deaths per capita is Louisiana. Mississippi, Tennessee, Alaska, and New Mexico also have a high death rate per capita. The state with the lowest gun deaths per capita is Hawaii. Other states such as Massachusetts, New Jersey, New York, and Rhode Island all have rates of less than 5 deaths per hundred thousand.
Events At The City/County Level
Below the US map chart is a table that shows examples of certain types of events such as mass shootings, school shootings, and accidental gun-related deaths at the city/county level. You can use the filtering described above to narrow down the states that you're interested in looking at which are then reflected in this more detailed table. Narrowing the search down to California, for example, lists shooting incidents in each local area within that state. You can use the scroll bar on the right-hand side of the table to page through incidents in each local area which are listed alphabetically.
We can see from these local statistics that, as expected, a larger number of incidents occur in local areas that are more populated. Also, more people appear to die in mass shootings than in accidental shootings.
Gun Law Effect Attribution
In an attempt to determine the effect that each gun law had on the state's gun deaths, we used a multi-step modeling process.
​
-
​We used NMF topic modeling from BERT embeddings to extract valuable information contained within gun law text and to represent the content of each gun law numerically.
-
Those numeric topic scores were placed into a Linear Mixed Effects model where the gun law topics were fixed effects that were dependent on the random effect of the state.
The output of the modeling process was a coefficient per gun law topic that varied depending on the state. This "effect score" represents the effect that a particular law had on a particular state over a period of one month. A negative value means that a particular gun law reduced gun deaths. A positive value means that a particular gun law increased gun deaths. We show the type of laws, whether the law is restrictive or permissive, and the full text of the provision by hovering over that portion of the text in the table.
​
We can select a specific state that we're interested in using the State selector on the upper right-hand corner of the table. Selecting Louisiana for example, it appears that two permissive laws may contribute to higher gun violence - the "castle doctrine" or "stand your ground" law and a law prohibiting local gun ordinances to be more restrictive than state laws. Waiting period provisions and background checks also have a negative correlation, which could cause us to consider evaluating those laws in more detail to determine their effectiveness.
CAUTION: It should be noted that this analytical method demonstrates potential CORRELATION between gun law provisions and gun law violence rates. Our methods do not establish CAUSATION. Further research would be needed further verify these results.
State Topic Recommendations
A "topic" refers to a specific subject or theme that is being discussed or addressed in a piece of text. In NLP, topic modeling is a popular technique used to identify and extract the main topics or themes from a collection of documents or a corpus of text.
Topic modeling algorithms aim to uncover hidden patterns in the text and group words or documents that are semantically related to each other. The output of a topic modeling process is a set of topics, where each topic is represented by a collection of words that are frequently co-occurring in the text.
​
Taking our analysis a step further, we can dive deeper into topic modeling. The visualization below represents a cutting-edge approach to understanding the impact of gun laws on gun-related deaths. We’ve taken the complex and varied gun laws across states and used advanced methods like Bidirectional Encoder Representations from Transformers (BERT) for tokenizing, PyTorch for extracting embeddings, and Non-Negative Matrix Factorization (NMF) for topic modeling, to distill these laws into 30 key topics. To make these topics human-readable, we employed Bidirectional Auto-Regressive Transformers (BART) and GPT-3.5 to construct understandable sentences. To assess the impact of each topic, we used a Linear Mixed Effect Model, a type of linear regression that accounts for variations between states and time periods. This model provides us with an estimate of the effect of implementing a particular topic in a state.
We then determined whether a topic was already implemented in a state by identifying the laws most likely to be associated with that topic. From the unimplemented topics, we used our model to identify those most likely to reduce gun deaths.
In essence, our visualization is a roadmap, guiding us toward the most effective legal changes to reduce gun violence in each state.
Conclusions and Recommendations
Evaluating the gun control provisions in states with the lowest occurrence of gun deaths, the following provisions appear to have potentially high impact on reducing gun deaths:
-
Background checks
-
Prohibited possessor laws (e.g. mental health, extreme risk protection)
-
Dealer license requirements
-
Restrictions on the minimum age
-
Required reporting of lost or stolen firearms
-
Child access laws related to negligent storage
Evaluating gun control provisions in states with the highest occurrence of gun deaths, it is unclear whether specific provisions are in common that have an impact on increasing gun deaths. However, it can clearly be observed that a variety of provisions such as those related to concealed weapons, minimum age, or prohibited possessor are considerably less detailed than provisions on similar topics in more restrictive states. Also, several states with higher gun deaths have permissive laws such as castle doctrine or open carry laws.
​
Looking at the top three recommendations across all states, the following were noted as the most common recommendations for states to improve their existing gun laws, in order of frequency:
-
Any person who sells, barters, gives, or furnishes a handgun to a minor, having good cause to believe them to be a minor, is guilty of a Class 6 felony and must undergo a criminal history background check
-
No person may carry a pistol or revolver, possess a dangerous weapon, firearm, or short-barreled shotgun, in any vehicle or on or about their person on school premises without a license or permit
-
No person or organization may sell, purchase, or repair firearms, destructive devices, parts, or ammunition without a dealer's license; registrants must comply with all duties imposed by this act
-
A licensing authority may grant a permit to purchase, rent, or lease firearms to adults, and the state has the authority to regulate firearms and knives, while municipalities may not
-
Minors under eighteen cannot possess or carry deadly weapons, except on their own or family premises or with parental/guardian permission
-
A protective order is a restraining order issued ex parte, after notice and hearing, or in a judgment, and prohibits a person subject to the order from owning, possessing, purchasing, or receiving firearms or ammunition while in effect
-
A firearms dealer must obtain a completed statement of intent, inspect proof of identity, and request a records check before transferring any handgun
​
Our conclusion is that certain gun laws do appear to have an impact on reducing gun violence while there are others either because of their content or because of the way they are written, have the effect of increasing gun violence. This tool can be a valuable tool for researchers and gun control advocates to analyze laws in specific states, how states compare to each other, and potential changes that can reduce gun violence. It is recommended that a more detailed evaluation be performed of the specific gun control text and associated implementation measures that have the most significant impact.