#######
#######
####### Replication files for Do Women Officers Police Differently? Evidence from Traffic Stops
####### This file produces the tables and figures seen in the paper and appendix.
####### Last Updated: Jan. 2021
#######
#######
###
### 1. Setting up the space.
###
# Setting the working directory:
setwd("~/Desktop/PinkPolicing/AJPS_ReplicationFiles")
# Installing the needed libraries:
#install.packages("pscl",dependencies = T)
#install.packages("ggplot2",dependencies = T)
#install.packages("texreg",dependencies = T)
#install.packages("readr",dependencies = T)
#install.packages("arm",dependencies = T)
# Opening up those libraries:
library(ggplot2)
library(texreg)
## Version: 1.37.5
## Date: 2020-06-17
## Author: Philip Leifeld (University of Essex)
##
## Consider submitting praise using the praise or praise_interactive functions.
## Please cite the JSS article in your publications -- see citation("texreg").
library(readr)
library(pscl)
## Classes and Methods for R developed in the
## Political Science Computational Laboratory
## Department of Political Science
## Stanford University
## Simon Jackman
## hurdle and zeroinfl functions by Achim Zeileis
library(arm)
## Loading required package: MASS
## Loading required package: Matrix
## Loading required package: lme4
##
## arm (Version 1.11-2, built: 2020-7-27)
## Working directory is /Users/kelseyshoub/Desktop/PinkPolicing/AJPS_ReplicationFiles
###
### 2. Body of the Paper
###
# Clearing the workspace + reading in data bit by bit to produce each table and figure.
rm(list = ls())
# Loading in the Data
load("Data/NorthCarolina.RData")
load("Data/FloridaLarge.RData")
load("Data/FloridaSmall.RData")
cmpd.employee = read_csv("Data/CMPD_Employee_Demographics.csv")
##
## ── Column specification ──────────────────────────────────────────────────────
## cols(
## JOB_TITLE = col_character(),
## Years_Of_Service = col_double(),
## Age = col_double(),
## Gender = col_character(),
## Race = col_character(),
## ObjectID = col_double()
## )
# Number of stops and searches by sex:
dim(fl)
## [1] 4842950 52
dim(nc)
## [1] 218158 32
table(fl$search_occur)
##
## 0 1
## 4391272 17356
table(nc$search)
##
## 0 1
## 207714 10444
prop.table(table(fl$search_occur))
##
## 0 1
## 0.996063174 0.003936826
prop.table(table(nc$search))
##
## 0 1
## 0.95212644 0.04787356
table(fl$of_gender)
##
## 0 1
## 3870641 291604
table(nc$of_gender)
##
## 0 1
## 199234 18924
table(fl$of_gender,fl$search_occur)
##
## 0 1
## 0 3843369 16422
## 1 290820 272
table(nc$of_gender,nc$search)
##
## 0 1
## 0 189611 9623
## 1 18103 821
prop.table(table(fl$of_gender,fl$search_occur),1)
##
## 0 1
## 0 0.9957453655 0.0042546345
## 1 0.9990655875 0.0009344125
prop.table(table(nc$of_gender,nc$search),1)
##
## 0 1
## 0 0.95170001 0.04829999
## 1 0.95661594 0.04338406
table(fl$of_gender,fl$contra)
##
## 0 1
## 0 3865730 4911
## 1 291491 113
# Number of officers by sex in FL
length(unique(fl$officer_id_hash))
## [1] 2708
length(unique(fl$officer_id_hash[fl$of_gender==0]))
## [1] 1916
length(unique(fl$officer_id_hash[fl$of_gender==1]))
## [1] 244
length(unique(fl$officer_id_hash[fl$officer_exclude==0]))
## [1] 2338
length(unique(fl$officer_id_hash[fl$of_gender==0&fl$officer_exclude==0]))
## [1] 1910
length(unique(fl$officer_id_hash[fl$of_gender==1&fl$officer_exclude==0]))
## [1] 244
table(cmpd.employee$JOB_TITLE[cmpd.employee$JOB_TITLE=="Police Officer"])
##
## Police Officer
## 1540
sum(table(cmpd.employee$Gender[cmpd.employee$JOB_TITLE=="Police Officer"]))
## [1] 1540
table(fl$year)
##
## 2010 2011 2012 2013 2014 2015
## 675487 870349 748026 830791 922008 796289
(table(fl$of_gender)/c(length(unique(fl$officer_id_hash[fl$of_gender==0&fl$officer_exclude==0])),length(unique(fl$officer_id_hash[fl$of_gender==1&fl$officer_exclude==0]))))/6
##
## 0 1
## 337.7523 199.1831
avg.stops = aggregate(fl$year,by=list(fl$officer_id_hash,fl$year,fl$of_gender),length)
summary(avg.stops)
## Group.1 Group.2 Group.3 x
## Length:9319 Length:9319 Min. :0.0000 Min. : 1.0
## Class :character Class :character 1st Qu.:0.0000 1st Qu.: 125.0
## Mode :character Mode :character Median :0.0000 Median : 359.0
## Mean :0.1062 Mean : 446.6
## 3rd Qu.:0.0000 3rd Qu.: 649.0
## Max. :1.0000 Max. :5299.0
mean(avg.stops$x)
## [1] 446.6407
median(avg.stops$x[avg.stops$Group.3==0])
## [1] 380
median(avg.stops$x[avg.stops$Group.3==1])
## [1] 202.5
prop.table(table(fl$investigatory[fl$of_gender==0]))
##
## 0 1
## 0.4649428 0.5350572
prop.table(table(fl$investigatory[fl$of_gender==1]))
##
## 0 1
## 0.4334131 0.5665869
table(nc$of_gender[nc$year==2019])[2:1]/table(cmpd.employee$Gender[cmpd.employee$JOB_TITLE=="Police Officer"])
##
## 1 0
## 33.13750 68.11154
# Excluding Cases:
dim(nc)
## [1] 218158 32
dim(nc)-dim(nc[!is.na(nc$search),])
## [1] 0 0
dim(fl)
## [1] 4842950 52
dim(fl)-dim(fl[!is.na(fl$search_occur),])
## [1] 434322 0
(dim(fl[!is.na(fl$search_occur),])-dim(fl.sm))+table(fl.sm$officer_exclude)[2]
## [1] 1695594 18170
table(fl.sm$county_include)
##
## 0 1
## 556 2730648
# Table 1
tab1 = data.frame("Department"=c("Charlotte PD (NC)",
"Male Officers","Female Officers",
"Florida Highwar Patrol",
"Male Officers","Female Officers"),
"Type"=c("Municipal","","","Statewide","",""),
"Years"=c("2016-2017","","",
"2010-2015","",""),
"Stops"=c(dim(nc)[1],table(nc$of_gender),
dim(fl[!is.na(fl$search_occur),])[1],
table(fl$of_gender[!is.na(fl$search_occur)])),
"Searches"=c(table(nc$search)[2],table(nc$of_gender,nc$search)[,2],
table(fl$search_occur)[2],
table(fl$of_gender,fl$search_occur)[,2]),
"Search Rate"=c(table(nc$search)[2]/dim(nc)[1],
table(nc$of_gender,nc$search)[,2]/table(nc$of_gender),
table(fl$search_occur)[2]/dim(fl[!is.na(fl$search_occur),])[1],
table(fl$of_gender,fl$search_occur)[,2]/
table(fl$of_gender[!is.na(fl$search_occur)])))
tab1 = rbind(tab1,
c("Total","","",
sum(tab1[c(1,4),4]),sum(tab1[c(1,4),5]),
sum(tab1[c(1,4),5])/sum(tab1[c(1,4),4])))
tab1
## Department Type Years Stops Searches
## 1 Charlotte PD (NC) Municipal 2016-2017 218158 10444
## 2 Male Officers 199234 9623
## 3 Female Officers 18924 821
## 4 Florida Highwar Patrol Statewide 2010-2015 4408628 17356
## 5 Male Officers 3859791 16422
## 6 Female Officers 291092 272
## 7 Total 4626786 27800
## Search.Rate
## 1 0.0478735595302487
## 2 0.048299988957708
## 3 0.0433840625660537
## 4 0.0039368256972464
## 5 0.00425463451259408
## 6 0.000934412488148077
## 7 0.00600849055910518
# Figure 1
load("Data/Fig1_Data.RData")
png("Figures/Fig1_PredProb.png",
750,519)
ggplot(data = search.df, aes(x=Department,y=Rate,fill=Gender)) +
geom_bar(stat="identity", position=position_dodge()) +
ylab("Search Rate") +
theme_bw(base_size=15)+
theme(legend.position = "bottom") +
labs(fill="Officer Sex")+
scale_fill_grey(start = 0.25, end = .75)
dev.off()
## quartz_off_screen
## 2
prop.test(table(fl$of_gender,fl$search_occur))
##
## 2-sample test for equality of proportions with continuity correction
##
## data: table(fl$of_gender, fl$search_occur)
## X-squared = 744.11, df = 1, p-value < 2.2e-16
## alternative hypothesis: two.sided
## 95 percent confidence interval:
## -0.003450662 -0.003189782
## sample estimates:
## prop 1 prop 2
## 0.9957454 0.9990656
prop.test(table(nc$of_gender,nc$search))
##
## 2-sample test for equality of proportions with continuity correction
##
## data: table(nc$of_gender, nc$search)
## X-squared = 9.0552, df = 1, p-value = 0.002619
## alternative hypothesis: two.sided
## 95 percent confidence interval:
## -0.007996240 -0.001835613
## sample estimates:
## prop 1 prop 2
## 0.9517000 0.9566159
# Table 2
load("Data/FLSearch_Sm_OLS.RData")
load("Data/FLSearch_OLS.RData")
load("Data/NCSearch_Sm_OLS.RData")
load("Data/NCSearch_OLS.RData")
screenreg(list(nc.search,fl.search),
stars=c(0.01,0.05),
custom.coef.map = list("(Intercept)"="(Intercept)",
"factor(of_gender)1"="Female Officer",
"factor(of_race)1"="Black Officer",
"factor(race_gender)1"="White Female",
"factor(race_gender)2"="Black Male",
"factor(race_gender)3"="Black Female",
"factor(race_gender)4"="Latino Male",
"factor(race_gender)5"="Latina Female",
"investigatory" = "Investigatory Stop Purpose"),
custom.model.names = c("(1) NC Search",
"(2) FL Search"),
digits=4)
##
## ===========================================================
## (1) NC Search (2) FL Search
## -----------------------------------------------------------
## (Intercept) 0.0862 ** 0.0263 **
## (0.0037) (0.0005)
## Female Officer -0.0256 ** -0.0038 **
## (0.0020) (0.0002)
## Black Officer -0.0292 ** -0.0028 **
## (0.0015) (0.0001)
## White Female -0.0086 ** -0.0026 **
## (0.0019) (0.0001)
## Black Male 0.0465 ** 0.0066 **
## (0.0016) (0.0001)
## Black Female -0.0204 ** -0.0015 **
## (0.0017) (0.0002)
## Latino Male 0.0015 **
## (0.0001)
## Latina Female -0.0020 **
## (0.0002)
## Investigatory Stop Purpose 0.0285 ** 0.0055 **
## (0.0012) (0.0001)
## -----------------------------------------------------------
## R^2 0.0713 0.0092
## Adj. R^2 0.0711 0.0091
## Num. obs. 150547 2712478
## ===========================================================
## ** p < 0.01; * p < 0.05
# Figure 2
fl.of.pred = predict(fl.search,
newdata = data.frame("of_gender"=c(0,1),"race_gender"=0,
"subject_age"=35,"out_of_state"=0,
"investigatory"=1,
"officer_years_of_service"=6,
"of_race"=0,"officer_age"=39,
"hour_of_day"=15,
"month"="05","year"=2013,
"county_name"="Orange County"),
type="response",se.fit=T)
nc.of.pred = predict(nc.search,
newdata = data.frame("of_gender"=c(0,1),
"race_gender"=0,
"subject_age"=36,
"investigatory"=1,
"Officer_Years_of_Service"=10.25,
"of_race"=0,"month"="01",
"year"=2019,"CMPD_Division"="South Division"),
type="response",se.fit=T)
pred.df = data.frame("Department" = c("Charlotte Police Department",
"Charlotte Police Department",
"Florida Highway Patrol",
"Florida Highway Patrol"),
"Gender" = c("Male","Female","Male","Female"),
"Predict" = c(nc.of.pred$fit,
fl.of.pred$fit),
"Lower"=c(nc.of.pred$fit-1.96*nc.of.pred$se.fit,
fl.of.pred$fit-1.96*fl.of.pred$se.fit),
"Upper"=c(nc.of.pred$fit+1.96*nc.of.pred$se.fit,
fl.of.pred$fit+1.96*fl.of.pred$se.fit))
png("Figures/Fig2_PredProb.png",
900,514)
ggplot(data = pred.df, aes(x=Gender,y=Predict)) +
geom_point(size=4) +
geom_errorbar(aes(ymin = Lower, ymax = Upper),
width=.2,size = 0.75,
position=position_dodge(.9)) +
ylab("Expected Probbility of a Search") +
xlab("Officer Sex") +
theme_bw(base_size=15) +facet_wrap(~Department)
dev.off()
## quartz_off_screen
## 2
pred.df$Predict[1]/pred.df$Predict[2]
## [1] 2.245964
pred.df$Predict[3]/pred.df$Predict[4]
## [1] 2.720538
# Table 3
tab3 = data.frame("Officer Gender"=c("Male","Female"),
"Searches"=table(fl$of_gender[!is.na(fl$search_occur)],
fl$search_occur[!is.na(fl$search_occur)])[,2],
"Contraband"=table(fl$of_gender[!is.na(fl$search_occur)],
fl$contra[!is.na(fl$search_occur)])[,2],
"Contraband Hit Rate"=table(fl$of_gender[!is.na(fl$search_occur)],
fl$contra[!is.na(fl$search_occur)])[,2]/
table(fl$of_gender[!is.na(fl$search_occur)],
fl$search_occur[!is.na(fl$search_occur)])[,2],
"Difference"=c((table(fl$of_gender[!is.na(fl$search_occur)],
fl$contra[!is.na(fl$search_occur)])[,2]/
table(fl$of_gender[!is.na(fl$search_occur)],
fl$search_occur[!is.na(fl$search_occur)])[,2])[1]-
(table(fl$of_gender[!is.na(fl$search_occur)],
fl$contra[!is.na(fl$search_occur)])[,2]/
table(fl$of_gender[!is.na(fl$search_occur)],
fl$search_occur[!is.na(fl$search_occur)])[,2])[2],NA))
tab3
## Officer.Gender Searches Contraband Contraband.Hit.Rate Difference
## 0 Male 16422 4911 0.2990501 -0.1163911
## 1 Female 272 113 0.4154412 NA
prop.test(table(fl$of_gender[fl$search_occur==1],
fl$contra[fl$search_occur==1]))
##
## 2-sample test for equality of proportions with continuity correction
##
## data: table(fl$of_gender[fl$search_occur == 1], fl$contra[fl$search_occur == 1])
## X-squared = 16.681, df = 1, p-value = 4.423e-05
## alternative hypothesis: two.sided
## 95 percent confidence interval:
## 0.05554105 0.17724120
## sample estimates:
## prop 1 prop 2
## 0.7009499 0.5845588
# Table 4
load("Data/FlContra_OLS.RData")
load("Data/FlSearchRate_OLS.RData")
load("Data/FlStopRate_OLS.RData")
screenreg(list(fl.contra,contra.search.rate.reg,contra.stop.rate.reg),
stars=c(0.01,0.05),
custom.coef.map = list("(Intercept)"="(Intercept)",
"factor(of_gender)1"="Female Officer",
"factor(of_race)1"="Black Officer",
"factor(race_gender)1"="White Female",
"factor(race_gender)2"="Black Male",
"factor(race_gender)3"="Black Female",
"factor(race_gender)4"="Latino Male",
"factor(race_gender)5"="Latina Female",
"investigatory" = "Investigatory Stop Purpose"),
custom.model.names = c("(1) Contra|Search",
"(2) Hit Rate, per 10 Searches",
"(3) Hit Rate, per 100 Stops"),
digits=4)
##
## =========================================================================================================
## (1) Contra|Search (2) Hit Rate, per 10 Searches (3) Hit Rate, per 100 Stops
## ---------------------------------------------------------------------------------------------------------
## (Intercept) 0.1118 ** 0.3006 0.1380 **
## (0.0421) (0.2148) (0.0176)
## Female Officer 0.1026 ** 1.1223 ** -0.0771 **
## (0.0294) (0.2760) (0.0117)
## Black Officer 0.0578 ** 0.7640 ** -0.0976 **
## (0.0199) (0.2030) (0.0096)
## White Female -0.0025 0.0512 -0.0557 **
## (0.0144) (0.1467) (0.0096)
## Black Male -0.0531 ** -0.4505 ** 0.0975 **
## (0.0097) (0.1058) (0.0104)
## Black Female -0.0594 ** -0.4565 ** -0.0519 **
## (0.0172) (0.1728) (0.0117)
## Latino Male -0.0909 ** -0.8755 ** -0.0021
## (0.0115) (0.1195) (0.0107)
## Latina Female -0.0027 0.0346 -0.0669 **
## (0.0267) (0.2586) (0.0128)
## Investigatory Stop Purpose 0.3394 ** 3.4794 ** 0.2534 **
## (0.0112) (0.1102) (0.0066)
## ---------------------------------------------------------------------------------------------------------
## R^2 0.1346 0.1311 0.0036
## Adj. R^2 0.1265 0.1285 0.0036
## Num. obs. 12782 9677 747784
## =========================================================================================================
## ** p < 0.01; * p < 0.05
###
### 3. Appendix A: Full Regression Results
###
screenreg(list(nc.search,fl.search,
fl.contra,contra.search.rate.reg,contra.stop.rate.reg),
stars=c(0.01,0.05),
custom.coef.map = list("(Intercept)"="(Intercept)",
"factor(of_gender)1"="Female Officer",
"factor(of_race)1"="Black Officer",
"officer_age"="Officer Age",
"factor(of_age)2"="Officer Age: 30-64",
"factor(of_age)3"="Officer Age: 65+",
"officer_years_of_service"="Officer Years of Service",
"Officer_Years_of_Service"="Officer Years of Service",
"factor(of_exper)1"="Experienced Officer",
"factor(race_gender)1"="White Female",
"factor(race_gender)2"="Black Male",
"factor(race_gender)3"="Black Female",
"factor(race_gender)4"="Latino Male",
"factor(race_gender)5"="Latina Female",
"subject_age"="Driver Age",
"factor(driver_age)2"="Driver Age: 30-64",
"factor(driver_age)3"="Driver Age: 65+",
"investigatory" = "Investigatory Stop Purpose",
"out_of_state"="Out of State"),
custom.model.names = c("(1)","(2)",
"(3)","(4)","(5)"),
digits=3)
##
## ===================================================================================================
## (1) (2) (3) (4) (5)
## ---------------------------------------------------------------------------------------------------
## (Intercept) 0.086 ** 0.026 ** 0.112 ** 0.301 0.138 **
## (0.004) (0.001) (0.042) (0.215) (0.018)
## Female Officer -0.026 ** -0.004 ** 0.103 ** 1.122 ** -0.077 **
## (0.002) (0.000) (0.029) (0.276) (0.012)
## Black Officer -0.029 ** -0.003 ** 0.058 ** 0.764 ** -0.098 **
## (0.001) (0.000) (0.020) (0.203) (0.010)
## Officer Age -0.000 ** -0.004 **
## (0.000) (0.001)
## Officer Age: 30-64 -0.375 ** -0.044 **
## (0.096) (0.008)
## Officer Age: 65+ -0.829 -0.262
## (4.048) (0.183)
## Officer Years of Service -0.002 ** 0.000 ** -0.000
## (0.000) (0.000) (0.001)
## Experienced Officer -0.026 0.053 **
## (0.086) (0.007)
## White Female -0.009 ** -0.003 ** -0.003 0.051 -0.056 **
## (0.002) (0.000) (0.014) (0.147) (0.010)
## Black Male 0.046 ** 0.007 ** -0.053 ** -0.451 ** 0.098 **
## (0.002) (0.000) (0.010) (0.106) (0.010)
## Black Female -0.020 ** -0.001 ** -0.059 ** -0.456 ** -0.052 **
## (0.002) (0.000) (0.017) (0.173) (0.012)
## Latino Male 0.001 ** -0.091 ** -0.876 ** -0.002
## (0.000) (0.011) (0.120) (0.011)
## Latina Female -0.002 ** -0.003 0.035 -0.067 **
## (0.000) (0.027) (0.259) (0.013)
## Driver Age -0.001 ** -0.000 ** -0.003 **
## (0.000) (0.000) (0.000)
## Driver Age: 30-64 -0.485 ** -0.123 **
## (0.085) (0.007)
## Driver Age: 65+ -1.113 * -0.187 **
## (0.446) (0.012)
## Investigatory Stop Purpose 0.028 ** 0.006 ** 0.339 ** 3.479 ** 0.253 **
## (0.001) (0.000) (0.011) (0.110) (0.007)
## Out of State 0.001 ** -0.053 ** -0.667 ** 0.037 **
## (0.000) (0.011) (0.110) (0.008)
## ---------------------------------------------------------------------------------------------------
## R^2 0.071 0.009 0.135 0.131 0.004
## Adj. R^2 0.071 0.009 0.127 0.128 0.004
## Num. obs. 150547 2712478 12782 9677 747784
## ===================================================================================================
## ** p < 0.01; * p < 0.05
###
### 4. Appendix B: Alternative Test of Differences in Search and Contraband Hit Rates
###
# Florida
fl$stop = 1
fl$of_exper = ifelse(fl$officer_years_of_service>=
mean(fl$officer_years_of_service,na.rm=T),1,0)
fl$of_age = ifelse(fl$officer_age<30,1,
ifelse(fl$officer_age>64,3,2))
fl$driver_age = ifelse(fl$subject_age<30,1,
ifelse(fl$subject_age>64,3,2))
fl$hour_of_day=as.numeric(fl$hour_of_day)
fl$tod = ifelse(fl$hour_of_day<3,1,
ifelse(fl$hour_of_day<6,2,
ifelse(fl$hour_of_day<9,3,
ifelse(fl$hour_of_day<12,4,
ifelse(fl$hour_of_day<15,5,
ifelse(fl$hour_of_day<18,6,
ifelse(fl$hour_of_day<21,7,8)))))))
fl.ag = aggregate(fl[!is.na(fl$search_occur),c("stop","search_occur","contra")],
by = list(fl$tod[!is.na(fl$search_occur)],
fl$officer_race[!is.na(fl$search_occur)],
fl$officer_sex[!is.na(fl$search_occur)],
fl$of_exper[!is.na(fl$search_occur)],
fl$race_gender[!is.na(fl$search_occur)],
fl$driver_age[!is.na(fl$search_occur)],
fl$out_of_state[!is.na(fl$search_occur)],
fl$investigatory[!is.na(fl$search_occur)]),
sum,na.rm=T)
colnames(fl.ag) = c("tod",
"of_race","of_sex","of_exper","driver_rg",
"driver_age","out_of_state","invest",
"stop","search","contraband")
fl.ag.female = fl.ag[fl.ag$of_sex=="female",]
colnames(fl.ag.female)[c(3,9:11)] = c("female","stop.f",
"search.f","contra.f")
fl.ag.male = fl.ag[fl.ag$of_sex=="male",]
colnames(fl.ag.male)[c(3,9:11)] = c("male","stop.m",
"search.m","contra.m")
fl.matches = merge(fl.ag.female,fl.ag.male)
min.stops = 9
table(fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops)
##
## FALSE TRUE
## 1461 1784
min.searches = 0
table(fl.matches$search.f>min.searches&
fl.matches$search.m>min.searches)
##
## FALSE TRUE
## 3084 161
table(fl.matches$search.f>min.searches&
fl.matches$search.m>min.searches&
fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops)
##
## FALSE TRUE
## 3084 161
# North Carolina
nc$stop = 1
nc$search = ifelse(nc$Was_a_Search_Conducted=="Yes",1,0)
nc$driver_age = ifelse(nc$Driver_Age<30,1,
ifelse(nc$Driver_Age>65,3,2))
nc$of_exper = ifelse(nc$Officer_Years_of_Service>=mean(nc$Officer_Years_of_Service),
1,0)
nc.ag = aggregate(nc[,c("search","stop")],
by = list(nc$CMPD_Division,
nc$Officer_Gender,nc$Officer_Race,
nc$of_exper,
nc$race_gender,nc$driver_age,
nc$investigatory,
nc$year),
sum)
nc.ag.female = nc.ag[nc.ag$Group.2=="Female",]
colnames(nc.ag.female) = c("division","female","race","of_exper",
"driver.rg","driver_age","investigatory",
"year",
"searches.f","stops.f")
nc.ag.male = nc.ag[nc.ag$Group.2=="Male",]
colnames(nc.ag.male) = c("division","male","race","of_exper",
"driver.rg","driver_age","investigatory",
"year",
"searches.m","stops.m")
# Searches
fl.matches$sr.f = fl.matches$search.f/fl.matches$stop.f
fl.matches$sr.m = fl.matches$search.m/fl.matches$stop.m
fl.matches$cr.f = fl.matches$contra.f/fl.matches$search.f
fl.matches$cr.m = fl.matches$contra.m/fl.matches$search.m
t.test(fl.matches$sr.f[fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops],
fl.matches$sr.m[fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops],
paired = T)
##
## Paired t-test
##
## data: fl.matches$sr.f[fl.matches$stop.f > min.stops & fl.matches$stop.m > min.stops] and fl.matches$sr.m[fl.matches$stop.f > min.stops & fl.matches$stop.m > min.stops]
## t = -13.359, df = 1783, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.003700686 -0.002753143
## sample estimates:
## mean of the differences
## -0.003226915
length(fl.matches$sr.f[fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops])
## [1] 1784
mean(fl.matches$sr.f[fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops])
## [1] 0.001355022
mean(fl.matches$sr.m[fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops])
## [1] 0.004581936
nc.matches = merge(nc.ag.female,nc.ag.male)
min.stops = 9
nc.matches$sr.f = nc.matches$searches.f/nc.matches$stops.f
nc.matches$sr.m = nc.matches$searches.m/nc.matches$stops.m
t.test(nc.matches$sr.f[nc.matches$stops.f>min.stops&
nc.matches$stops.m>min.stops],
nc.matches$sr.m[nc.matches$stops.f>min.stops&
nc.matches$stops.m>min.stops],
paired = T)
##
## Paired t-test
##
## data: nc.matches$sr.f[nc.matches$stops.f > min.stops & nc.matches$stops.m > min.stops] and nc.matches$sr.m[nc.matches$stops.f > min.stops & nc.matches$stops.m > min.stops]
## t = -4.0127, df = 352, p-value = 7.335e-05
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.024123468 -0.008254384
## sample estimates:
## mean of the differences
## -0.01618893
length(nc.matches$sr.f[nc.matches$stops.f>min.stops&
nc.matches$stops.m>min.stops])
## [1] 353
mean(nc.matches$sr.f[nc.matches$stops.f>min.stops&
nc.matches$stops.m>min.stops])
## [1] 0.05538775
mean(nc.matches$sr.m[nc.matches$stops.f>min.stops&
nc.matches$stops.m>min.stops],)
## [1] 0.07157667
# Contraband
t.test(fl.matches$cr.f[fl.matches$search.f>min.searches&
fl.matches$search.m>min.searches&
fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops],
fl.matches$cr.m[fl.matches$search.f>min.searches&
fl.matches$search.m>min.searches&
fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops],
paired = T)
##
## Paired t-test
##
## data: fl.matches$cr.f[fl.matches$search.f > min.searches & fl.matches$search.m > min.searches & fl.matches$stop.f > min.stops & fl.matches$stop.m > min.stops] and fl.matches$cr.m[fl.matches$search.f > min.searches & fl.matches$search.m > min.searches & fl.matches$stop.f > min.stops & fl.matches$stop.m > min.stops]
## t = 2.6679, df = 160, p-value = 0.008419
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## 0.02401357 0.16088431
## sample estimates:
## mean of the differences
## 0.09244894
length(fl.matches$cr.f[fl.matches$search.f>min.searches&
fl.matches$search.m>min.searches&
fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops])
## [1] 161
mean(fl.matches$cr.f[fl.matches$search.f>min.searches&
fl.matches$search.m>min.searches&
fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops])
## [1] 0.4090506
mean(fl.matches$cr.m[fl.matches$search.f>min.searches&
fl.matches$search.m>min.searches&
fl.matches$stop.f>min.stops&
fl.matches$stop.m>min.stops])
## [1] 0.3166016
###
### 5. Appendix C: Logistic Regrssion Models
###
rm(list = ls())
load("Data/FlContra_Logit.RData")
load("Data/FLSearch_Logit.RData")
load("Data/NCSearch_Logit.RData")
texreg(list(nc.search,fl.search,fl.contra),
stars=c(0.01,0.05),
custom.coef.map = list("(Intercept)"="(Intercept)",
"factor(of_gender)1"="Female Officer",
"factor(of_race)1"="Black Officer",
"factor(of_race)2"="Latinx Officer",
"factor(of_race)3"="Asain/Pacific Islander Officer",
"factor(of_race)4"="Other Race Officer",
"officer_age"="Officer Age",
"officer_years_of_service"="Officer Years of Service",
"Officer_Years_of_Service"="Officer Years of Service",
"factor(race_gender)1"="White Female",
"factor(race_gender)2"="Black Male",
"factor(race_gender)3"="Black Female",
"factor(race_gender)4"="Latino Male",
"factor(race_gender)5"="Latina Female",
"subject_age"="Driver Age",
"investigatory" = "Investigatory Stop Purpose",
"out_of_state"="Out of State"),
custom.model.names = c("(1) NC Search",
"(2) FL Search",
"(3) FL Contra|Search"),
digits=4)
##
## \begin{table}
## \begin{center}
## \begin{tabular}{l c c c}
## \hline
## & (1) NC Search & (2) FL Search & (3) FL Contra|Search \\
## \hline
## (Intercept) & $-1.9244^{**}$ & $-2.8175^{**}$ & $-17.9811$ \\
## & $(0.0906)$ & $(0.0900)$ & $(148.7052)$ \\
## Female Officer & $-0.4702^{**}$ & $-1.4253^{**}$ & $0.4986^{**}$ \\
## & $(0.0477)$ & $(0.0674)$ & $(0.1547)$ \\
## Black Officer & $-0.7213^{**}$ & $-1.0929^{**}$ & $0.2932^{**}$ \\
## & $(0.0387)$ & $(0.0441)$ & $(0.1060)$ \\
## Latinx Officer & & $-0.3512^{**}$ & $-0.1044$ \\
## & & $(0.0363)$ & $(0.0924)$ \\
## Asain/Pacific Islander Officer & & $-1.0206^{**}$ & $0.7392$ \\
## & & $(0.2019)$ & $(0.5612)$ \\
## Other Race Officer & & $-0.8249^{**}$ & $0.8317^{*}$ \\
## & & $(0.1632)$ & $(0.3917)$ \\
## Officer Age & & $-0.0221^{**}$ & $-0.0202^{**}$ \\
## & & $(0.0012)$ & $(0.0036)$ \\
## Officer Years of Service & $-0.0770^{**}$ & $0.0129^{**}$ & $-0.0002$ \\
## & $(0.0023)$ & $(0.0017)$ & $(0.0045)$ \\
## White Female & $-0.6166^{**}$ & $-0.8276^{**}$ & $-0.0151$ \\
## & $(0.0751)$ & $(0.0337)$ & $(0.0782)$ \\
## Black Male & $0.8877^{**}$ & $0.8883^{**}$ & $-0.2900^{**}$ \\
## & $(0.0436)$ & $(0.0228)$ & $(0.0534)$ \\
## Black Female & $-0.4638^{**}$ & $-0.2870^{**}$ & $-0.3230^{**}$ \\
## & $(0.0518)$ & $(0.0404)$ & $(0.0960)$ \\
## Latino Male & & $0.3641^{**}$ & $-0.5462^{**}$ \\
## & & $(0.0276)$ & $(0.0663)$ \\
## Latina Female & & $-0.7432^{**}$ & $0.0039$ \\
## & & $(0.0624)$ & $(0.1465)$ \\
## Driver Age & $-0.0422^{**}$ & $-0.0450^{**}$ & $-0.0161^{**}$ \\
## & $(0.0012)$ & $(0.0008)$ & $(0.0022)$ \\
## Investigatory Stop Purpose & $0.6995^{**}$ & $1.5916^{**}$ & $17.9495$ \\
## & $(0.0298)$ & $(0.0262)$ & $(148.7051)$ \\
## Out of State & & $0.3653^{**}$ & $-0.3378^{**}$ \\
## & & $(0.0269)$ & $(0.0667)$ \\
## \hline
## AIC & $49914.2052$ & $137958.1571$ & $13507.4833$ \\
## BIC & $50261.4763$ & $139508.5753$ & $14402.1785$ \\
## Log Likelihood & $-24922.1026$ & $-68858.0786$ & $-6633.7416$ \\
## Deviance & $49844.2052$ & $137716.1571$ & $13267.4833$ \\
## Num. obs. & $150547$ & $2712478$ & $12782$ \\
## \hline
## \multicolumn{4}{l}{\scriptsize{$^{**}p<0.01$; $^{*}p<0.05$}}
## \end{tabular}
## \caption{Statistical models}
## \label{table:coefficients}
## \end{center}
## \end{table}
###
### 6. Appendix C: Fixed Effects
###
rm(list = ls())
load("Data/FLSearch_OLS_FE.RData")
load("Data/FlContra_OLS_FE.RData")
load("Data/FlSearchRate_OLS_FE.RData")
load("Data/FlStopRate_OLS_FE.RData")
texreg(list(fl.search,
fl.contra,
contra.search.rate.reg,
contra.stop.rate.reg),
stars=c(0.01,0.05),
custom.coef.map = list("(Intercept)"="(Intercept)",
"factor(of_gender)1"="Female Officer",
"factor(of_race)1"="Black Officer",
"officer_age"="Officer Age",
"factor(of_age)2"="Officer Age: 30-64",
"factor(of_age)3"="Officer Age: 65+",
"officer_years_of_service"="Officer Years of Service",
"Officer_Years_of_Service"="Officer Years of Service",
"factor(of_exper)1"="Experienced Officer",
"factor(race_gender)1"="White Female",
"factor(race_gender)2"="Black Male",
"factor(race_gender)3"="Black Female",
"factor(race_gender)4"="Latino Male",
"factor(race_gender)5"="Latina Female",
"subject_age"="Driver Age",
"factor(driver_age)2"="Driver Age: 30-64",
"factor(driver_age)3"="Driver Age: 65+",
"investigatory" = "Investigatory Stop Purpose",
"out_of_state"="Out of State"),
custom.model.names = c("(1) Search",
"(2) Contra|Search",
"(3) Hit Rate, per 10 Searches",
"(4) Hit Rate, per 100 Stops"),
digits=4)
##
## \begin{table}
## \begin{center}
## \begin{tabular}{l c c c c}
## \hline
## & (1) Search & (2) Contra|Search & (3) Hit Rate, per 10 Searches & (4) Hit Rate, per 100 Stops \\
## \hline
## (Intercept) & $0.0142^{**}$ & $0.1363^{*}$ & $0.2596$ & $0.0707^{**}$ \\
## & $(0.0017)$ & $(0.0567)$ & $(0.2384)$ & $(0.0232)$ \\
## Female Officer & $-0.0026^{*}$ & $0.0916^{*}$ & $0.9424^{*}$ & $-0.0531$ \\
## & $(0.0013)$ & $(0.0376)$ & $(0.3793)$ & $(0.0334)$ \\
## Black Officer & $-0.0035^{**}$ & $0.0541$ & $0.3139$ & $-0.0894^{**}$ \\
## & $(0.0012)$ & $(0.0399)$ & $(0.4081)$ & $(0.0306)$ \\
## Officer Age & $-0.0001^{**}$ & $-0.0048^{**}$ & & \\
## & $(0.0000)$ & $(0.0012)$ & & \\
## Officer Age: 30-64 & & & $-0.4058^{**}$ & $-0.0231$ \\
## & & & $(0.1474)$ & $(0.0160)$ \\
## Officer Age: 65+ & & & $-0.5371$ & $-0.1241$ \\
## & & & $(4.0924)$ & $(0.2141)$ \\
## Officer Years of Service & $0.0002^{**}$ & $0.0024$ & & \\
## & $(0.0000)$ & $(0.0016)$ & & \\
## Experienced Officer & & & $0.3207^{*}$ & $0.0692^{**}$ \\
## & & & $(0.1614)$ & $(0.0151)$ \\
## White Female & $-0.0022^{**}$ & $-0.0037$ & $0.0647$ & $-0.0545^{**}$ \\
## & $(0.0001)$ & $(0.0142)$ & $(0.1418)$ & $(0.0096)$ \\
## Black Male & $0.0051^{**}$ & $-0.0535^{**}$ & $-0.5417^{**}$ & $0.0851^{**}$ \\
## & $(0.0001)$ & $(0.0097)$ & $(0.1042)$ & $(0.0104)$ \\
## Black Female & $-0.0019^{**}$ & $-0.0609^{**}$ & $-0.4922^{**}$ & $-0.0630^{**}$ \\
## & $(0.0002)$ & $(0.0170)$ & $(0.1682)$ & $(0.0116)$ \\
## Latino Male & $0.0013^{**}$ & $-0.0909^{**}$ & $-0.8668^{**}$ & $-0.0088$ \\
## & $(0.0001)$ & $(0.0114)$ & $(0.1177)$ & $(0.0107)$ \\
## Latina Female & $-0.0019^{**}$ & $-0.0114$ & $-0.1267$ & $-0.0675^{**}$ \\
## & $(0.0002)$ & $(0.0264)$ & $(0.2512)$ & $(0.0128)$ \\
## Driver Age & $-0.0001^{**}$ & $-0.0023^{**}$ & & \\
## & $(0.0000)$ & $(0.0004)$ & & \\
## Driver Age: 30-64 & & & $-0.3525^{**}$ & $-0.1190^{**}$ \\
## & & & $(0.0829)$ & $(0.0069)$ \\
## Driver Age: 65+ & & & $-0.8633^{*}$ & $-0.1688^{**}$ \\
## & & & $(0.4308)$ & $(0.0117)$ \\
## Investigatory Stop Purpose & $0.0041^{**}$ & $0.3340^{**}$ & $3.3725^{**}$ & $0.2428^{**}$ \\
## & $(0.0001)$ & $(0.0112)$ & $(0.1081)$ & $(0.0067)$ \\
## Out of State & $0.0018^{**}$ & $-0.0544^{**}$ & $-0.5205^{**}$ & $0.0317^{**}$ \\
## & $(0.0001)$ & $(0.0112)$ & $(0.1093)$ & $(0.0083)$ \\
## \hline
## AIC & $-6993843.0503$ & $15022.3545$ & $54081.7043$ & $3674955.2660$ \\
## BIC & $-6992267.0054$ & $15931.9613$ & $54311.3845$ & $3675324.0618$ \\
## Log Likelihood & $3497044.5251$ & $-7389.1773$ & $-27008.8521$ & $-1837445.6330$ \\
## Num. obs. & $2712478$ & $12782$ & $9677$ & $747784$ \\
## Num. groups: officer\_id\_hash & $1419$ & $599$ & $$ & $$ \\
## Var: officer\_id\_hash (Intercept) & $0.0002$ & $0.0150$ & $$ & $$ \\
## Var: Residual & $0.0044$ & $0.1744$ & $14.8268$ & $7.9441$ \\
## Num. groups: officer\_id & $$ & $$ & $602$ & $1424$ \\
## Var: officer\_id (Intercept) & $$ & $$ & $2.1381$ & $0.1270$ \\
## \hline
## \multicolumn{5}{l}{\scriptsize{$^{**}p<0.01$; $^{*}p<0.05$}}
## \end{tabular}
## \caption{Statistical models}
## \label{table:coefficients}
## \end{center}
## \end{table}
###
### 7. Appendix D: Interaction Models
###
rm(list = ls())
# Table 1. Officer Experience
load("Data/FLSearch_Exper_OLS.RData")
load("Data/NCSearch_Exper_OLS.RData")
load("Data/FlContra_Exper_OLS.RData")
load("Data/FlSearchRate_Exper_OLS.RData")
load("Data/FlStopRate_Exper_OLS.RData")
texreg(list(nc.search.exper,fl.search.exper,fl.contra.exper,
contra.search.rate.exper,contra.stop.rate.exper),
stars=c(0.05,0.01),
custom.coef.map = list("factor(of_gender)1"="Female Officer",
"officer_years_of_service"="Officer Years of Service",
"Officer_Years_of_Service"="Officer Years of Service",
"factor(of_exper)1"="Experienced Officer",
"factor(of_gender)1:officer_years_of_service"="Female Officer * Exper.",
"factor(of_gender)1:Officer_Years_of_Service"="Female Officer * Exper.",
"factor(of_gender)1:factor(of_exper)1"="Female Officer * Exper."),
digits = 3)
##
## \begin{table}
## \begin{center}
## \begin{tabular}{l c c c c c}
## \hline
## & Model 1 & Model 2 & Model 3 & Model 4 & Model 5 \\
## \hline
## Female Officer & $-0.025^{**}$ & $-0.004^{**}$ & $0.112^{**}$ & $0.971^{**}$ & $-0.059^{**}$ \\
## & $(0.003)$ & $(0.000)$ & $(0.040)$ & $(0.364)$ & $(0.015)$ \\
## Officer Years of Service & $-0.002^{**}$ & $0.000^{**}$ & $-0.000$ & & \\
## & $(0.000)$ & $(0.000)$ & $(0.001)$ & & \\
## Experienced Officer & & & & $-0.043$ & $0.056^{**}$ \\
## & & & & $(0.087)$ & $(0.008)$ \\
## Female Officer * Exper. & $-0.000$ & $-0.000$ & $-0.002$ & $0.443$ & $-0.047^{*}$ \\
## & $(0.000)$ & $(0.000)$ & $(0.005)$ & $(0.558)$ & $(0.024)$ \\
## \hline
## R$^2$ & $0.071$ & $0.009$ & $0.135$ & $0.129$ & $0.003$ \\
## Adj. R$^2$ & $0.071$ & $0.009$ & $0.126$ & $0.127$ & $0.003$ \\
## Num. obs. & $150547$ & $2712478$ & $12782$ & $9677$ & $747784$ \\
## \hline
## \multicolumn{6}{l}{\scriptsize{$^{**}p<0.01$; $^{*}p<0.05$}}
## \end{tabular}
## \caption{Statistical models}
## \label{table:coefficients}
## \end{center}
## \end{table}
# Table 2. Prop Female
load("Data/FLSearch_Prop_OLS.RData")
load("Data/FlContra_Prop_OLS.RData")
texreg(list(fl.search.prop,fl.contra.prop),
stars=c(0.05,0.01),
custom.coef.map = list("factor(of_gender)1"="Female Officer",
"female.prop"="Female Proportion of Proximate Force",
"factor(of_gender)1:female.prop"="Female Officer * Female Prop."),
digits = 3)
##
## \begin{table}
## \begin{center}
## \begin{tabular}{l c c}
## \hline
## & Model 1 & Model 2 \\
## \hline
## Female Officer & $-0.003^{**}$ & $0.434^{**}$ \\
## & $(0.001)$ & $(0.105)$ \\
## Female Proportion of Proximate Force & $-0.004$ & $-0.269$ \\
## & $(0.002)$ & $(0.203)$ \\
## Female Officer * Female Prop. & $-0.010$ & $-3.350^{**}$ \\
## & $(0.006)$ & $(1.020)$ \\
## \hline
## R$^2$ & $0.009$ & $0.136$ \\
## Adj. R$^2$ & $0.009$ & $0.127$ \\
## Num. obs. & $2712478$ & $12782$ \\
## \hline
## \multicolumn{3}{l}{\scriptsize{$^{**}p<0.01$; $^{*}p<0.05$}}
## \end{tabular}
## \caption{Statistical models}
## \label{table:coefficients}
## \end{center}
## \end{table}
# Table 3. Stop Type
load("Data/FLSearch_StopType_OLS.RData")
load("Data/NCSearch_StopType_OLS.RData")
load("Data/FlContra_StopType_OLS.RData")
load("Data/FlSearchRate_StopType_OLS.RData")
load("Data/FlStopRate_StopType_OLS.RData")
texreg(list(nc.search.st,fl.search.st,fl.contra.st,
contra.search.rate.st,contra.stop.rate.st),
stars=c(0.05,0.01),
custom.coef.map = list("(Intercept)"="(Intercept)",
"factor(of_gender)1"="Female Officer",
"factor(of_race)1"="Black Officer",
"officer_age"="Officer Age",
"factor(of_age)2"="Officer Age: 30-64",
"factor(of_age)3"="Officer Age: 65+",
"officer_years_of_service"="Officer Years of Service",
"Officer_Years_of_Service"="Officer Years of Service",
"factor(of_exper)1"="Experienced Officer",
"factor(race_gender)1"="White Female",
"factor(race_gender)2"="Black Male",
"factor(race_gender)3"="Black Female",
"factor(race_gender)4"="Latino Male",
"factor(race_gender)5"="Latina Female",
"subject_age"="Driver Age",
"factor(driver_age)2"="Driver Age: 30-64",
"factor(driver_age)3"="Driver Age: 65+",
"investigatory" = "Investigatory Stop Purpose",
"out_of_state"="Out of State"),
digits = 3)
##
## \begin{table}
## \begin{center}
## \begin{tabular}{l c c c c c}
## \hline
## & Model 1 & Model 2 & Model 3 & Model 4 & Model 5 \\
## \hline
## (Intercept) & $0.145^{**}$ & $0.045^{**}$ & $0.485^{**}$ & $3.754^{**}$ & $0.449^{**}$ \\
## & $(0.006)$ & $(0.001)$ & $(0.048)$ & $(0.170)$ & $(0.025)$ \\
## Female Officer & $-0.032^{**}$ & $-0.006^{**}$ & $0.114^{**}$ & $1.322^{**}$ & $-0.147^{**}$ \\
## & $(0.003)$ & $(0.000)$ & $(0.034)$ & $(0.323)$ & $(0.022)$ \\
## Black Officer & $-0.039^{**}$ & $-0.005^{**}$ & $0.065^{**}$ & $0.892^{**}$ & $-0.195^{**}$ \\
## & $(0.002)$ & $(0.000)$ & $(0.023)$ & $(0.242)$ & $(0.019)$ \\
## Officer Age & & $-0.000^{**}$ & $-0.004^{**}$ & & \\
## & & $(0.000)$ & $(0.001)$ & & \\
## Officer Age: 30-64 & & & & $-0.495^{**}$ & $-0.096^{**}$ \\
## & & & & $(0.115)$ & $(0.016)$ \\
## Officer Age: 65+ & & & & & $-0.436$ \\
## & & & & & $(0.310)$ \\
## Officer Years of Service & $-0.003^{**}$ & $0.000^{**}$ & $-0.000$ & & \\
## & $(0.000)$ & $(0.000)$ & $(0.001)$ & & \\
## Experienced Officer & & & & $-0.037$ & $0.103^{**}$ \\
## & & & & $(0.104)$ & $(0.014)$ \\
## White Female & $-0.018^{**}$ & $-0.004^{**}$ & $-0.004$ & $0.068$ & $-0.115^{**}$ \\
## & $(0.004)$ & $(0.000)$ & $(0.017)$ & $(0.174)$ & $(0.019)$ \\
## Black Male & $0.055^{**}$ & $0.010^{**}$ & $-0.061^{**}$ & $-0.558^{**}$ & $0.184^{**}$ \\
## & $(0.003)$ & $(0.000)$ & $(0.011)$ & $(0.127)$ & $(0.020)$ \\
## Black Female & $-0.028^{**}$ & $-0.002^{**}$ & $-0.068^{**}$ & $-0.562^{**}$ & $-0.105^{**}$ \\
## & $(0.003)$ & $(0.000)$ & $(0.020)$ & $(0.208)$ & $(0.023)$ \\
## Latino Male & & $0.002^{**}$ & $-0.109^{**}$ & $-1.087^{**}$ & $-0.004$ \\
## & & $(0.000)$ & $(0.013)$ & $(0.143)$ & $(0.021)$ \\
## Latina Female & & $-0.003^{**}$ & $0.005$ & $0.091$ & $-0.135^{**}$ \\
## & & $(0.000)$ & $(0.032)$ & $(0.314)$ & $(0.025)$ \\
## Driver Age & $-0.002^{**}$ & $-0.000^{**}$ & $-0.003^{**}$ & & \\
## & $(0.000)$ & $(0.000)$ & $(0.000)$ & & \\
## Driver Age: 30-64 & & & & $-0.630^{**}$ & $-0.244^{**}$ \\
## & & & & $(0.101)$ & $(0.014)$ \\
## Driver Age: 65+ & & & & $-1.573^{**}$ & $-0.398^{**}$ \\
## & & & & $(0.567)$ & $(0.023)$ \\
## Out of State & & $0.003^{**}$ & $-0.065^{**}$ & $-0.864^{**}$ & $0.066^{**}$ \\
## & & $(0.000)$ & $(0.013)$ & $(0.133)$ & $(0.016)$ \\
## \hline
## R$^2$ & $0.071$ & $0.012$ & $0.084$ & $0.047$ & $0.003$ \\
## Adj. R$^2$ & $0.070$ & $0.012$ & $0.074$ & $0.045$ & $0.003$ \\
## Num. obs. & $79523$ & $1474530$ & $11041$ & $8045$ & $382456$ \\
## \hline
## \multicolumn{6}{l}{\scriptsize{$^{**}p<0.01$; $^{*}p<0.05$}}
## \end{tabular}
## \caption{Statistical models}
## \label{table:coefficients}
## \end{center}
## \end{table}
# Table 4. Driver Characteristics
load("Data/FLInter_Search.RData")
load("Data/FLInter_Contra.RData")
load("Data/FLStopRate_Inter_OLS.RData")
load("Data/FLSearchRate_Inter_OLS.RData")
load("Data/NCInter_Search.RData")
texreg(list(nc.search.inter,fl.search.inter,fl.contra.inter,
contra.search.rate.inter,contra.stop.rate.inter),
stars=c(0.01,0.05),
custom.coef.map = list("factor(of_gender)1"="Female Officer",
"factor(subject_female)1"="Female Driver",
"factor(of_race)1"="Black Officer",
"factor(of_race)2"="Latinx Officer",
"factor(subject_race2)1"="Black Driver",
"factor(subject_race2)2"="Latinx Driver",
"factor(of_gender)1:factor(subject_female)1"="Female Officer*Driver",
"factor(of_race)1:factor(subject_race2)1"="Black Officer*Driver",
"factor(of_race)2:factor(subject_race2)1"="Latinx Officer*Black Driver",
"factor(of_race)1:factor(subject_race2)2"="Black Officer*Latinx Driver",
"factor(of_race)2:factor(subject_race2)2"="Latinx Officer* Driver"),digits=3)
##
## \begin{table}
## \begin{center}
## \begin{tabular}{l c c c c c}
## \hline
## & Model 1 & Model 2 & Model 3 & Model 4 & Model 5 \\
## \hline
## Female Officer & $-0.024^{**}$ & $-0.005^{**}$ & $0.111^{**}$ & $1.216^{**}$ & $-0.098^{**}$ \\
## & $(0.002)$ & $(0.000)$ & $(0.034)$ & $(0.315)$ & $(0.015)$ \\
## Female Driver & $-0.046^{**}$ & $-0.004^{**}$ & $0.010$ & $0.956^{**}$ & $-0.071^{**}$ \\
## & $(0.001)$ & $(0.000)$ & $(0.010)$ & $(0.266)$ & $(0.014)$ \\
## Black Officer & $-0.017^{**}$ & $-0.001^{**}$ & $0.020$ & $0.415$ & $-0.077^{**}$ \\
## & $(0.002)$ & $(0.000)$ & $(0.031)$ & $(0.330)$ & $(0.014)$ \\
## Latinx Officer & $-0.020^{**}$ & $-0.000$ & $0.063^{*}$ & $1.338^{**}$ & $-0.032^{*}$ \\
## & $(0.005)$ & $(0.000)$ & $(0.026)$ & $(0.247)$ & $(0.015)$ \\
## Black Driver & $0.027^{**}$ & $0.006^{**}$ & $-0.044^{**}$ & $-1.295^{**}$ & $0.035$ \\
## & $(0.001)$ & $(0.000)$ & $(0.009)$ & $(0.318)$ & $(0.019)$ \\
## Latinx Driver & $-0.007^{**}$ & $0.002^{**}$ & $-0.075^{**}$ & $-0.800^{**}$ & $0.004$ \\
## & $(0.002)$ & $(0.000)$ & $(0.012)$ & $(0.129)$ & $(0.012)$ \\
## Female Officer*Driver & $0.003$ & $0.003^{**}$ & $-0.042$ & $-0.315$ & $0.046$ \\
## & $(0.004)$ & $(0.000)$ & $(0.068)$ & $(0.641)$ & $(0.024)$ \\
## Black Officer*Driver & $-0.018^{**}$ & $-0.005^{**}$ & $0.085^{*}$ & $0.790$ & $-0.056^{*}$ \\
## & $(0.003)$ & $(0.000)$ & $(0.043)$ & $(0.470)$ & $(0.023)$ \\
## Latinx Officer*Black Driver & $-0.001$ & $-0.003^{**}$ & $-0.167^{**}$ & $-1.922^{**}$ & $-0.113^{**}$ \\
## & $(0.006)$ & $(0.000)$ & $(0.033)$ & $(0.350)$ & $(0.025)$ \\
## Black Officer*Latinx Driver & $0.002$ & $-0.002^{**}$ & $0.018$ & $0.219$ & $-0.016$ \\
## & $(0.005)$ & $(0.000)$ & $(0.047)$ & $(0.513)$ & $(0.024)$ \\
## Latinx Officer* Driver & $0.012$ & $-0.002^{**}$ & $-0.088^{**}$ & $-1.132^{**}$ & $-0.038$ \\
## & $(0.008)$ & $(0.000)$ & $(0.034)$ & $(0.348)$ & $(0.023)$ \\
## \hline
## R$^2$ & $0.063$ & $0.009$ & $0.132$ & $0.132$ & $0.004$ \\
## Adj. R$^2$ & $0.063$ & $0.009$ & $0.124$ & $0.130$ & $0.003$ \\
## Num. obs. & $176332$ & $2658706$ & $12718$ & $9677$ & $747784$ \\
## \hline
## \multicolumn{6}{l}{\scriptsize{$^{**}p<0.01$; $^{*}p<0.05$}}
## \end{tabular}
## \caption{Statistical models}
## \label{table:coefficients}
## \end{center}
## \end{table}
###
### 8. Appendix E: A Conservative Test with the Charlotte Police Department
###
load("Data/NorthCarolina.RData")
table(nc$year)
##
## 2016 2017 2019 2020
## 41113 46943 96498 33604
nc.search16 = lm(search~factor(race_gender)+subject_age+
investigatory+
factor(of_race)+
factor(of_gender)+Officer_Years_of_Service+
factor(month)+
factor(CMPD_Division),
data=nc,subset=nc$year==2016)
nc.search17 = lm(search~factor(race_gender)+subject_age+
investigatory+
factor(of_race)+
factor(of_gender)+Officer_Years_of_Service+
factor(month)+
factor(CMPD_Division),
data=nc,subset=nc$year==2017)
nc.search19 = lm(search~factor(race_gender)+subject_age+
investigatory+
factor(of_race)+
factor(of_gender)+Officer_Years_of_Service+
factor(month)+
factor(CMPD_Division),
data=nc,subset=nc$year==2019)
nc.search20 = lm(search~factor(race_gender)+subject_age+
investigatory+
factor(of_race)+
factor(of_gender)+Officer_Years_of_Service+
factor(month)+
factor(CMPD_Division),
data=nc,subset=nc$year==2020)
texreg(list(nc.search16,nc.search17,nc.search19,nc.search20),
omit.coef = "Division*|month*",
custom.coef.map = list("(Intercept)"="(Intercept)",
"factor(of_gender)1"="Female Officer",
"factor(of_race)1"="Black Officer",
"Officer_Years_of_Service"="Officer Years of Service",
"investigatory"="Investigatory Stop",
"factor(race_gender)1"="White Female",
"factor(race_gender)2"="Black Male",
"factor(race_gender)3"="Black Female",
"subject_age"="Driver Age"),
stars=c(0.01,0.05))
##
## \begin{table}
## \begin{center}
## \begin{tabular}{l c c c c}
## \hline
## & Model 1 & Model 2 & Model 3 & Model 4 \\
## \hline
## (Intercept) & $0.10^{**}$ & $0.09^{**}$ & $0.09^{**}$ & $0.10^{**}$ \\
## & $(0.01)$ & $(0.01)$ & $(0.01)$ & $(0.01)$ \\
## Female Officer & $-0.03^{**}$ & $-0.03^{**}$ & $-0.02^{**}$ & $-0.02^{**}$ \\
## & $(0.00)$ & $(0.00)$ & $(0.00)$ & $(0.01)$ \\
## Black Officer & $-0.03^{**}$ & $-0.03^{**}$ & $-0.03^{**}$ & $-0.04^{**}$ \\
## & $(0.00)$ & $(0.00)$ & $(0.00)$ & $(0.00)$ \\
## Officer Years of Service & $-0.00^{**}$ & $-0.00^{**}$ & $-0.00^{**}$ & $-0.00^{**}$ \\
## & $(0.00)$ & $(0.00)$ & $(0.00)$ & $(0.00)$ \\
## Investigatory Stop & $0.02^{**}$ & $0.02^{**}$ & $0.03^{**}$ & $0.04^{**}$ \\
## & $(0.00)$ & $(0.00)$ & $(0.00)$ & $(0.00)$ \\
## White Female & $-0.01^{**}$ & $-0.00$ & $-0.01^{**}$ & $-0.01$ \\
## & $(0.00)$ & $(0.00)$ & $(0.00)$ & $(0.01)$ \\
## Black Male & $0.04^{**}$ & $0.05^{**}$ & $0.04^{**}$ & $0.04^{**}$ \\
## & $(0.00)$ & $(0.00)$ & $(0.00)$ & $(0.00)$ \\
## Black Female & $-0.02^{**}$ & $-0.01^{**}$ & $-0.02^{**}$ & $-0.03^{**}$ \\
## & $(0.00)$ & $(0.00)$ & $(0.00)$ & $(0.01)$ \\
## Driver Age & $-0.00^{**}$ & $-0.00^{**}$ & $-0.00^{**}$ & $-0.00^{**}$ \\
## & $(0.00)$ & $(0.00)$ & $(0.00)$ & $(0.00)$ \\
## \hline
## R$^2$ & $0.07$ & $0.06$ & $0.08$ & $0.09$ \\
## Adj. R$^2$ & $0.07$ & $0.06$ & $0.08$ & $0.09$ \\
## Num. obs. & $31275$ & $34701$ & $64501$ & $20070$ \\
## \hline
## \multicolumn{5}{l}{\scriptsize{$^{**}p<0.01$; $^{*}p<0.05$}}
## \end{tabular}
## \caption{Statistical models}
## \label{table:coefficients}
## \end{center}
## \end{table}