Who fires a gun at a school? Data science provides eight archetypes of shooters
In product marketing, a “customer archetype” is a common set of attributes, interests, and behaviors that represent a segment of customers.
How did Netflix get so good at predicting what movies and TV shows you like? In 2009, Netflix offered $1 million to whoever could create the best algorithm for predicting customer preferences based on a dataset of 1.4 million customer ratings.
The model that won the contest was able to accurately predict what a viewer would like 86% of the time by using a data science technique called collaborative filtering. Identifying likenesses in groups of users is such a powerful tool that Netflix had to shut down the project due to concerns the customer data could predict a wide array of attributes beyond TV preferences.
In product marketing, a “customer archetype” is a common set of attributes, interests, and behaviors that represent a segment of customers. The idealized stereotypes go beyond basic demographics and purchase history by identifying groups that share similar patterns. Netflix can predict that you will like David Chang’s new cooking show because other users with your archetype watched multiple episodes of it.
Applicability to gun violence at schools
To better understand who fires a gun on a k-12 school campus, I created a model that used the data from 2,700 shootings at schools since 1966 to identify eight different archetypes of a school shooter.
Unlike millions of datapoints that Netflix can use to model future behaviors, there is no predictive model and not enough data available to identify a specific kid who is plotting an attack. The utility of these school shooter archetypes is they allow school officials and police to plan for different sets of circumstances without needing to know the identity of an attacker.
It’s also important to note that these archetypes are not real school shooters. Any real person—past, present, or future—can be a combination of multiple archetypes.
Here are the results from my data analysis model:
1. The Adolescent Insider
School Affiliation: Student
Age: Teen
Gender: Male (predominantly)
Typical Outcome: Fled/Apprehended
Victim Count: Usually 1 or fewer indicating incidents might be targeted.
Situation: This archetype could be involved in a variety of situations, but Escalation of Dispute, reflecting arguments or conflicts that turn violent, and Bullying, indicating a response to being bullied, might be particularly relevant. Additionally, Anger Over Grade/Suspension/Discipline could reflect incidents motivated by academic or disciplinary grievances.
Time Period: Likely during Morning Classes or Afternoon Classes, reflecting the regular school day when students are present.
Location: Classrooms, Hallways, or Cafeterias where students spend most of their time.
Likely Weapon Type: Handgun, given its accessibility and ease of concealment. This choice might reflect opportunistic or impulsive actions rather than extensive planning.
Profile: A student deeply embedded within the school community, often acting out of impulsive or emotional distress. They typically use a handgun, reflecting accessibility and a lack of extensive planning. Their actions, occurring during regular school hours in common areas, stem from personal grievances or social pressures.
Keep reading with a 7-day free trial
Subscribe to School Shooting Data Analysis and Reports to keep reading this post and get 7 days of free access to the full post archives.