Utilizing the data file from the previous assignment and the listed variables, complete the following:
Variable list:
race_coded
sex
threat_coded
attack_dummy
flight
Mental_illness
body_cam
weapon
firearm
replica
Alabama
Age
Age_cat
Provide descriptive statistics for these variables.
Conduct a correlation
What is the correlation coefficient for each of the listed variables with (1) threat_coded, (2) attack_dummy, and (3) mental_illness.
Are any of these significant? If so which ones?
Conduct a simple regression with the following and provide whether the x variable was significant and the coefficient:
sex (x), threat_coded (y)
race_coded (x), threat_coded (y)
mental_illness (x), attack (y)
Conduct multivariate analysis (OLS) with the following models and provide the x variables that were significant and their coefficient:
(x) All variables excluding attack_dummy, replica, firearm, and age. (y) threat_coded
Do the same as above, but use age instead of age_coded
How did the results change?
What variables were significant in each of these models?
How much did these models explain the variance in the threat associated with police related fatalities (R-squared)?
(x) all variables excluding threat_coded, replica, firearm, and age_coded. (y) attack_dummy
How did the results change between the previous model using age_cat (1.1)?
(x) all variables excluding threat_coded, weapon, and age_coded. (y) attack_dummy
Conduct the same as above but exclude the variable Alabama
What changes were observed?
How much did these models explain the variance in during attacks resulting in police related fatalities (R-squared)?
(x) All variables. (y) attack_dummy
Did you run into any issues running this model?
If so, why do you think this is?
Across all models (bivariate and multivariate) describe and detail any reoccurring findings. To rephrase, what do these findings suggest about threats and attacks resulting in police related fatalities?
Notes:
Remember your variables must be independent. As such, you cannot use your y variable as an x variable in the model. Makle sure that you are not including your dependent variable (y) among the independent variables (y).
Interpretation (this is an explanation of what you are seeing with these data)
for threat_coded – these results indicate as the level of threat increases, the influence of (variable-name) in the suspect’s/citizen’s (insert term) death increased/decreased. (Since we are using a categorical outcome variable there is little that can be derived from this outside of using specific analysis not covered in this course)
for attack_dummy – these results indicate that during incidents where an attack occurred, (variable-name) the influence of (variable-name) in the suspect’s/citizen’s (insert term) death increased/decreased.
This/These (variables) contributed/did not contribute (recall all cases resulted in death) to suspects deaths in officer involved fatalities.
