| MAX_TACTIC_USED | Mean Positive Emotionality | Mean Negative Emotionality | Mean Contstraint |
|---|---|---|---|
| Firearm | 56.48 | 47.02 | 49.20 |
| No Tactic Used | 56.46 | 46.15 | 48.72 |
| Pepper spray | 54.37 | 45.75 | 46.74 |
| Taser | 55.96 | 46.14 | 47.11 |
| Verbal descalaiton | 56.86 | 47.12 | 48.23 |
Police Use of Violence and Personality
Background
This data, from the National Institute of Justice, contains data gathered from 64 law enforcement officers and 101 civilians. The original work outlines an attempt to alter citizen views on police officers, as well as demonstrating relationships between personality measures and behavior in a officer training simulation. Data gathered from this study include a battery of survey responses and observations of participants’ tactics while completing an officer training simulation. See here for a full account of data.
The original work outlines an attempt to alter citizen views on police officers, as well as demonstrating relationships between personality measures and behavior in a officer training simulation.
This project:
As this project evolves through the semester, I plan to explore the multi-dimensional personality questionnaire brief form’s (MPQ-BF) relationship to behaviors in a training simulation. Here, initial exploration reveals a potential relationship between sub factors of negative emotionality and violent behaviors. I want to extend this exploration to other variables in the future. For example: the simulation scenario (low, med, high risk) and survey responses regarding officer social network as it pertains to the MPQ-BF sub-factor, Alienation.
Research Questions:
Does the multi dimensional personality questionnaire brief form correlate with behavior?
In other words, is there evidence for criterion validity?
Do higher traits of certain personality traits show increased violence in samples of police officers?
Data Preparation
A data frame was created with the following variables
Data included:
-Participant ID Number
-Sample: officer or civilian
-MPQ-BF survey results :155 questions with three primary factors:
1.Positive Emotionality: constructed from, wellbeing, social potency, and achievement
2.Negative Emotinality: which consists of stress reaction, alienation, and aggression
3.Constraint: composed of control, harm avoidance, and traditionalism.
-SCENARIO:Firearms Training Scenario
-N_DEESCALATE_TOTAL:Total number of verbal de-escalation statements/actions a subject uses within the scenario
-N_FIREARM_TOTAL: Total number of times a subject uses a firearm within the scenario
-N_SPRAY_TOTAL: Total number of times a subject uses pepper spray within scenario
-N_TASER_TOTAL: Total number of times a subject uses a taser within scenario
-MAX_TACTIC_USED: Categorical variable for most severe use of force used by the subject during the scenario
-N_FORCE_BY_YOU: (officer only) response to question: Over the course of your career as a law enforcement officer (across all departments), how many times -if any- have you used lethal force against another person in the line of duty?
First Manipulation
Participant sample, weapon used, and officer use of force recoded to numeric values.
Next, a new column totaling instances of weapon (firearm, taser, pepper spray) use by participant.
Lastly, the creation of seperate civilian and officer data frames to allow for easier descriptive analyses
Civilian Descriptive Statistics
Mean MPQ scores group by the weapon participant used
Civilians MPQ score by how many times they used a weapon during the simulation
| total_weapon_used | Mean Positive Emotionality | Mean Negative Emotionality | Mean Contstraint |
|---|---|---|---|
| 0 | 56.56 | 46.39 | 48.60 |
| 1 | 56.12 | 46.95 | 48.34 |
| 2 | 56.23 | 43.79 | 45.45 |
| 3 | 46.89 | 40.51 | 51.52 |
Officer Descriptive Statistics
Officer Mean MPQ Grouped by Weapon Used in Simulation.
| MAX_TACTIC_USED | Mean Positive Emotionality | Mean Negative Emotionality | Mean Contstraint |
|---|---|---|---|
| Firearm | 57.56 | 47.10 | 50.31 |
| No Tactic Used | 56.96 | 46.06 | 52.58 |
| Pepper spray | 52.70 | 47.18 | 45.05 |
| Taser | 56.27 | 46.74 | 50.61 |
| Verbal descalaiton | 57.90 | 45.83 | 52.41 |
Officer MPQ score by how many times they used a weapon during the simulation
| total_weapon_used | Mean Positive Emotionality | Mean Negative Emotionality | Mean Contstraint |
|---|---|---|---|
| 0 | 57.59 | 45.90 | 52.47 |
| 1 | 56.97 | 46.77 | 50.34 |
| 2 | 55.43 | 50.43 | 46.49 |
| 3 | 63.29 | 42.55 | 56.03 |
Officer Mean MPQ grouped by the amount of times officer has used deadly force on the job
| Numeric_Force | Mean Positive Emotionality | Mean Negative Emotionality | Mean Constraint |
|---|---|---|---|
| more than once | 55.09 | 50.55 | 54.22 |
| once | 58.74 | 50.31 | 48.09 |
| zero | 57.14 | 45.45 | 51.35 |
Additional manipulation to Officer Data Frame
A new variable is created combining real and simulated uses of violence. With one instance catergorized as low, two as medium, and three or more as high. These categories are created to keep group sizes similar. See descriptive statistics of this new variable below.
| High_LOW_sim_real | Mean Positive Emotionality | Mean Negative Emotionality | Mean Constraint |
|---|---|---|---|
| high | 57.72 | 49.44 | 50.92 |
| low | 58.33 | 44.88 | 52.92 |
| medium | 56.08 | 46.41 | 49.82 |
The above table and plot showing average MPQ scores by instances of both simulated and real violence shows a higher mean in negative emotion. Therefore, the sub factors of negative emotionality (aggression, stress, and alienation) are examined further.
Negative Emotionality sub-factor exploration
| High_LOW_sim_real | Mean Aggression | Mean Stress Reaction | Mean Alienation |
|---|---|---|---|
| high | 49.32 | 44.80 | 53.18 |
| low | 43.78 | 41.52 | 50.94 |
| medium | 46.47 | 42.23 | 51.92 |
Plot Demonstrating above Descriptive Statistics
Results
Overall, this exploration shows potential statistically significant relationships between the MPQ-BF and behaviors within the simulation:
For civilians, the use of fire arms is related to higher negative emotionality, compared to less lethal weapons like pepper spray or a taser. Additionally, high constraint and low positive emotionality seem to relate to increased use of weapons. Oddly, the mean for civilian negative emotionality was low for those who more often used weapons in the simulation.
Similar relationships are also seen with the Officer sample, with those use weapons scoring slightly higher in negative emotionality than those who only used verbal deescalation . Oddly enough, officers scoring high in Positive emotionality more frequently used weapons, and those who used many weapons had lower negative emotionality.
However, when examining real-world instances of an officer using lethal force, an expected negative relationship between negative emotionality and instances of violence resumes. Additionlly, we can see a relationship between nonviolence and positive emotionality.
Reflection
I learned a lot from this assignment. One of the largest takeaways was the practice I gained writing syntax from memory. I did my best not to look up syntax on the internet unless I really needed to, as I felt this might make it hard for me to memorize syntax I use a lot. Instead, I predominantly relied on notes and examples from class. I feel as though this helped me contextualize the information we’ve learned so far.
Another example of what I learned is the importance of carefully inspecting codebook when you are downloading a dataset. There was an instance where use of force by polic officers was coded as 1 for 0, 2 for one, and so on. I should have seen this sooner.
Additionally, gaining insight into all variables, and planning an approach is crucial. With R, you need to think about the objective you want to complete before you can make any progress. There needs to be an expressed goal–you have to aim at something. I gained a lot of insight into both syntax and the data set; I am excited to continue refining my plan.