Look Twice and Save Data Chanyoung Lee Ph

Look Twice and Save Data Chanyoung Lee, Ph. D. Program Director, Motorcycle Injury Prevention Institute (MIPI) Center for Urban Transportation Research (CUTR)

Amateur vs. Pro SMSA 2015 NATIONAL SYMPOSIUM

Motorcycle data collection and analysis Gear Use (Helmet) Endorsement/ Licensing Registration Crash VMT Training SMSA 2015 NATIONAL SYMPOSIUM

Motorcycle Helmet SMSA 2015 NATIONAL SYMPOSIUM

How to Measure the Use of Helmet � Discusses the advantages and disadvantages of three common data collection methods for motorcycle helmet use • • • Stated-Preference (SP) survey Crash data Observational survey SMSA 2015 NATIONAL SYMPOSIUM

Which of the following best describes your use of a motorcycle helmet? SMSA 2015 NATIONAL SYMPOSIUM

Endorsed Motorcyclists in Florida SMSA 2015 NATIONAL SYMPOSIUM

Simple Random Sampling (SRS) 40. 0% 35. 0% 30. 0% 24. 6% 25. 0% 20. 0% 13. 9% 15. 0% 23. 7% 17. 1% 16. 4% 10. 0% 5. 0% 1. 0% 3. 3% 0. 0% 18 -20 21 -24 Survey Respondents 25 -34 35 -44 45 -54 Endorsed Riders in Florida 55 -64 65 or older Motorcycle Crash SMSA 2015 NATIONAL SYMPOSIUM

Strong Statistical Association SMSA 2015 NATIONAL SYMPOSIUM

Age vs. VMT SMSA 2015 NATIONAL SYMPOSIUM

Annual Motorcycle VMT Estimate in Florida (Based on median value in 2015 Florida Motorcyclists survey) 7000 6665 6000 5000 4780 4734 4934 25 -34 35 -44 45 -54 4924 55 -64 5147 4000 3000 2000 1000 0 21 -24 65 or older SMSA 2015 NATIONAL SYMPOSIUM

Stated-Preference (SP) Surveys � Low response � Sampling • Motorcycle Endorsement or Registration � Difficult • to quantify 80 trips/100 trips vs. 90 trips/100 trips � Cognitive • rate (Survey Fatigue) bias can also be introduced easily Reliability of self-response data is questionable for socially undesirable behaviors SMSA 2015 NATIONAL SYMPOSIUM

We all do better Controlling the motorcycle (i e vehicle control skills): In your opinion, how do you ride compared to other riders of your age and gender? Staying sober while riding a motorcycle: In your opinion, how do you ride compared to other riders of your age and gender? SMSA 2015 NATIONAL SYMPOSIUM

Survey Data http: //www. scoopwhoop. com/world/hoc-quotes/ SMSA 2015 NATIONAL SYMPOSIUM

Stated-Preference (SP) Surveys � Can • collect a broad range of information Behavioral characteristics SMSA 2015 NATIONAL SYMPOSIUM

Crash Data SMSA 2015 NATIONAL SYMPOSIUM

Florida Traffic Crash Report SMSA 2015 NATIONAL SYMPOSIUM

Motorcycle Crash Data in Florida PPE Information Collected in FDOT Crash Reports More Specific Information Collected Up to October 2010 After October 2010 Safety Equipment Use Helmet Use (HU) 1 Not in Use 1 DOT-Compliant Motorcycle 2 Seat Belt/Shoulder Harness Helmet 3 Child Restraint 2 Other Helmet 4 Air Bag – Deployed 3 No Helmet 5 Air Bag – Not Deployed Eye Protection (EP) 6 Safety Helmet 1 Yes 7 Eye Protection 2 No 3 Not Applicable SMSA 2015 NATIONAL SYMPOSIUM

Helmet Use by Month in FL 70% 90 80 70 65% 60% % of Helmeted Motorcycle Riders 60 Avg. Temperature (°F) 50 40 30 55% 20 10 50% 0 1 2 3 4 5 6 7 Month 8 9 10 11 12 SMSA 2015 NATIONAL SYMPOSIUM

Motorcycle crash data analysis � Surrogate measure � Systematic reporting errors due to inconsistent practices in the field and other problems � Motorcycle Type • Sport Bike (Almost 80%) vs. Cruiser (40%) � Large Sample � Low cost Size - 24/7/365 days SMSA 2015 NATIONAL SYMPOSIUM

Observational Surveys � Less bias or error � Sufficient sample size needed/Costly SMSA 2015 NATIONAL SYMPOSIUM

Sample Size Matters http: //www-nrd. nhtsa. dot. gov/Pubs/812010. pdf SMSA 2015 NATIONAL SYMPOSIUM

Many States Collect Helmet Use Information through Observational Surveys � Most of these surveys have been conducted in conjunction with a statewide seatbelt use survey � As a result, the methodology largely followed the Uniform Criteria for State Observational Surveys of Seat Belt Use from Title 23, Part 1240. 12 of the Code of Federal Regulations • Sample Size/Day of week/Roadways SMSA 2015 NATIONAL SYMPOSIUM

Observed Use of Motorcycle Helmets in Florida Year DOTCompliant Non. No Compliant Helmet # of Observations 2010 52. 4% 1. 3% 46. 3% 5, 196 2011 49. 2% 3. 4% 47. 4% 7, 547 2012 47. 0% 3. 1% 49. 9% 10, 363 2013 50. 7% 2. 9% 46. 4% 9, 464 2014* 49. 3% 3. 3% 47. 3% 7, 642 Avg. 49. 7% 2. 8% 47. 5% 8, 042 *Excluding Monroe county observations SMSA 2015 NATIONAL SYMPOSIUM

Observed Helmet Use by Motorcycle Type All DOTCompliant Cruiser Custom Moped/ On/Off Scooter Road Sport Bike Standard Touring Trike 41. 2% 19. 1% 16. 0% 52. 9% 79. 6% 59. 6% 48. 0% 48. 9% 32. 3% 13. 6% 8. 6% 5. 7% 1. 7% 21. 2% 37. 9% 43. 0% Full Face 8. 9% 5. 5% 7. 2% 37. 1% 77. 9% 37. 0% 10. 0% 5. 9% Motocross 0. 0% 0. 2% 10. 0% 1. 4% 0. 0% Noncompliant 4. 7% 5. 5% 1. 2% 1. 4% 0. 4% 3. 2% 4. 7% 2. 6% All Unhelmeted 54. 1% 75. 5% 82. 8% 45. 7% 20. 0% 37. 2% 47. 3% 48. 5% No Helmet 38. 2% 54. 5% 68. 2% 41. 4% 16. 3% 29. 2% 31. 4% 32. 9% Decorative 15. 3% 20. 9% 14. 6% 4. 3% 2. 3% 7. 7% 15. 6% 14. 7% Carrying 0. 5% 0. 0% 1. 4% 0. 3% 1. 0% Sample Size 3011 110 2115 70 1420 349 1616 307 Open Face SMSA 2015 NATIONAL SYMPOSIUM

PPE use in Florida Winter (Jan/Feb) and Summer (May/June) Armor Vest 1. 0% 2. 9% 5. 8% 8. 4% 93. 6% 98. 3% Closed Toe Shoes 69. 6% Long Pants 21. 0% Gloves Jacket 7. 0% 39. 6% 52. 4% 98. 2% 98. 6% Eye-Protection Non-Compliant Helmet 90. 8% 5. 2% 2. 6% 44. 3% 54. 2% DOT-Compliant Helmet 0. 0%10. 0%20. 0%30. 0%40. 0%50. 0%60. 0%70. 0%80. 0%90. 0% 100. 0% 110. 0% Summer (May/June) Winter (Jan/Feb) SMSA 2015 NATIONAL SYMPOSIUM

Issues � Resource � Sampling • issues Week of Day, County, Motorcycle Type � Quality of observer SMSA 2015 NATIONAL SYMPOSIUM

Metric Accuracy for estimating helmet use Stated-Preference (SP) Crash Data Analysis Surveys Self-response bias can be Includes only crashed introduced motorcycle riders; can be vulnerable to systematic errors if any data is missing data Observational Surveys In general, high accuracy expected; data quality can be subject to sample size and data collection method, including training of data collection personnel Comprehensiveness Collects a range of information, such rider demographics, characteristics, attitudes, and preferences in conjunction with helmet use Collects a reasonable amount of information, such as gender and age; attitudes and preferences not available Collects observable characteristics (helmet use, motorcycle type); limited to collect unobservable information; training experience, age, etc. Cost High Low-cost if state collects and stores helmet information in crash data High Feasibility of longitudinal data Timeliness Possible Can collect data and analyze in 1– 2 months Varies. 1 -24 months Can collect data and analyze in. SYMPOSIUM 1– 2 months SMSA 2015 NATIONAL

Look Twice and Save Data SMSA 2015 NATIONAL SYMPOSIUM

Conclusion � Motorcycle Safety is a challenging issue � “Data-Driven” � It is not a bad idea will require resources � Bias, Common sense, Hidden layer � “All models are wrong but some are useful” George Edward Pelham Box SMSA 2015 NATIONAL SYMPOSIUM

Any Questions or Comments? Chanyoung Lee cylee@cutr. usf. edu 813. 974. 5307