This study analyzed the effects of ridesharing accessibility on DUI arrests in a university city using statistical approaches. Weekly and daily regression models were conducted using DUI arrest data from 2010-2015 and controlling for variables like football games, ridesharing programs, and semester. The results suggested a marginal effect of Uber and ridesharing programs in reducing DUI arrests. However, additional variables are needed to better quantify ridesharing demand and effects on DUIs.
1. USING STATISTICAL APPROACHES TO QUANTIFY THE EFFECTS OF RIDESHARING ACCESSIBILITY ON
DRIVING UNDER THE INFLUENCE (DUI) ARRESTS IN A UNIVERSITY CITY
Sheldon Waugh, MS and Jacob Ball, MA
Thursday, October 20, 2016
Department of Epidemiology
College of Public Health and Health Professions
University of Florida
2. Presenter Disclosures
(1) The following personal financial relationships with
commercial interests relevant to this presentation
existed during the past 12 months:
Sheldon Waugh
Jacob Ball
No relationships to disclose
3. DUIs and University Culture
• 1,825 college students between
the ages of 18 and 24 die from
alcohol-related unintentional
injuries, including motor-vehicle
crashes 1.
• At age 19, 17% of students drove
while intoxicated, 42% drove after
drinking any alcohol, and 38%
rode with an intoxicated driver 2.
0
10
20
30
40
50
60
18 19 20 21 22 Male Female
Percent
Binge Alcohol Use in the Past Month among Persons Aged
18 to 22, by College Enrollment Status and Demographic
Characteristics (2013)
Full Time College Students Other Persons
4. DUIs and University Culture (College Football)
• Student college football fans may
represent a significant risk factor
for binge drinking, with fans
reporting higher alcohol
consumption on game days than
non fan students3.
• Football game days were
associated with increased alcohol
consumption and higher number
of alcohol related arrests2-4.
5. Ridesharing: a good alternative?
• Technologies that utilize smartphones to
arrange temporary shared rides in real time.
• Popular ridesharing companies use contracted
drivers and spatial supply and demand
framework to match ride requesters with
drivers.
• Other positives:
• Cheaper than taxis
• Higher safety and quality of service
• Quicker response times
6. Uber Safe Rides Program
• Originated in 2014
• Enacted to increase use of safe modes of
transportation at night
• Uber offered cooperative partnerships with
large universities
• Students with active student ID numbers
offered a 50 percent discount for Uber rides
within a certain area, including campus
• Offered nighttime hours (10pm - 3am)
• Paid with university subsidies and student
government transportation fee
7. What’s Missing? - Question - Rationale
Little has been done to quantify the effects of these interventions in terms of
reducing DUIs arrests. Additionally, significant city events such as college football
must be taken in to consideration.
• What is the effect of ridesharing on DUI arrests in a university city?
• What, if any, effect did the ridesharing subsidy program have on DUI arrests?
• To develop a model to appropriately quantify the effects of ridesharing programs
combined with mediators and covariates (College Football) known to affect DUI
arrests in a university city.
• Hypotheses:
• Introduction of ridesharing : ↓ DUI Arrests
• Introduction of Safe Rides Program : marginal ↓ DUI Arrests
8. Data
• Arrest records and logs were collected from the University police
department and city police department
• All logged DUI arrests (Driving Under the Influence) were tagged with initial
date and time recorded along with unique report ID. DUI data were obtained
from January of 2010 to 2016.
• Unique SafeRides Program discount requests from students at the
university
• Unique discount code requests were logged and collected by a Southeastern
University participating in the collaborative program, from the beginning of
the trial period (April 2015) to December of 2015.
13. Methods
• A 5 year (2010-2015) weekly Poisson regression model
• Response variable: Weekly Counts of DUI Arrests
• Variables: Year, Presence of Football, Interaction term (Football and Year), School
Semester, Quintile cumulative discount downloads, Presence of Uber (July 2014)
• A 2 year daily Hurdle regression model (2014-2015)
• Hurdle regression analysis which employs a logistic regression for the
presence/absence of weeks with at least one DUI, and then a truncated Poisson or
negative binomial regression to predict the magnitude of DUIs given there is at least
one in a given week.
• Response variable: Daily Counts of DUI Arrests
• Variables: Weekday, Year, Month, College football game site, Quintile cumulative
discount downloads, Presence of Uber (July 2014) Arrests only from Wednesday to
Saturday are included in the model.
• OR/IRR > 1: Adverse effect
• OR/IRR < 1: Protective effect
15. Uber Safe Rides
Football Game Site
Year
Controlling for:
Weekday, Year, Month,
Daily Hurdle Regression
16. Conclusions
• Marginal effect of ridesharing
• Introduction of Safe Rides Program : marginal ↓ DUI Arrests
• Additional variables are necessary to improve the quantification of
ridesharing demand and use.
• Surge (Uber)
• Police presence may be another variable to observe in future studies.
17. References
1. Hingson, R. W., Zha, W. & Weitzman, E. R. Magnitude of and Trends in Alcohol-Related
Mortality and Morbidity Among U.S. College Students Ages 18-24, 1998-2005. J Stud
Alcohol Drugs Suppl 12–20 (2009).
2. Beck, K. H. et al. Trends in alcohol-related traffic risk behaviors among college
students. Alcohol. Clin. Exp. Res. 34, 1472–1478 (2010).
3. Tavis Glassman MPH, Mse., PhD;, C. E. W., Edessa Jobli MD, M. & PhD, H. B. Alcohol-
Related Fan Behavior on College Football Game Day. Journal of American College
Health 56, 255–260 (2007).
4. Merlo, L. J., Hong, J. & Cottler, L. B. The association between alcohol-related arrests
and college football game days. Drug and Alcohol Dependence 106, 69–71 (2010).
5. Merlo, L. J., Ahmedani, B. K., Barondess, D. A., Bohnert, K. M. & Gold, M. S. Alcohol
consumption associated with collegiate American football pre-game festivities. Drug
and Alcohol Dependence 116, 242–245 (2011).
6. Rees, D. I. & Schnepel, K. T. College Football Games and Crime. Journal of Sports
Economics 10, 68–87 (2009).