Estimating the Total Number of Propeller Accidents: A Senior Design Project
Safety professionals turn to the U.S. Coast Guard Boating Accident Reports Database (BARD) to find the total number of recreational boat propeller accidents reported to USCG.
USCG acknowledges some propeller accidents go unreported. The boating industry claims the more severe an accident it is, the more likely it is to be reported, propeller accidents are severe, so they must almost all be reported.
Propeller safety activists point to countless unreported accidents and previous studies showing thousands of propeller accidents going unreported.
Without an estimate of the total number of propeller accidents, injuries, and fatalities it is difficult to decide how best to address the problem. This same problem exists in accident frequency studies in automobile crashes, industrial accidents, and other fields. We propose encouraging students searching for Senior Design Projects, Capstone Projects, Sr. Thesis, Masters Thesis, and similar projects to consider applying the techniques used to estimate actual accident frequencies in other fields (like automobile injuries and fatalities) to recreational boat propeller accidents.
We will begin by supplying some information on previous under reporting of propeller accident studies, discussing techniques used to estimate the actual number of accidents in other fields of study, describing the accidents currently NOT included in USCG’s BARD database, and conclude by supplying some references.
Previous Studies of Under Reporting of Propeller Accidents
We cite several previous studies on our Accident Statistics page in the Under Reporting of Accidents section.
How Other Fields Have Estimated the Number of Unreported Accidents
There are five basic strategies. Three of these strategies have previously been applied to recreational boat propeller accidents.
Previously Used Strategies
- Increase the percentage of accidents being reported – capture reports at hospitals, tougher laws requiring them to be reported, incentives for reporting, extra layers of people searching for unreported accidents and making sure they get reported, etc. For example, for the last few years, USCG has hired a contractor capturing boat accident reports from the media. If the states do not report a particular accident, USCG sends them the media clip and asks them to follow up.
- Apply extra resources to try to attempt to force reporting frequencies to 100 percent in one or more small geographical areas. Compare the near 100 percent data to the previous data, estimate the percent of non reported accidents, extrapolate that to the region or to the entire country. The Center for Disease Control and others have taken this approach for propeller accidents.
- Telephone Surveys – call thousands of participants and ask them if they have been involved in an accident of the type under study during the previous year. The Red Cross did this for propeller accidents during 1988-1989 boating season.
To our knowledge, the remaining two strategies have not been applied to recreational boat propeller injuries. These are the strategies we are proposing for Senior Design Projects, Capstone Projects, Masters Thesis and similar projects.
- Compare the results of two independent databases of accidents, then use statistical techniques based on the overlap (the percentage of accidents reported in both accident databases) to estimate the total number of accidents.
- Compare the ratios/percentages of accidents, injury frequencies, and fatality frequencies across time or across regions, countries, etc. For example, if two countries with relatively similar geography, populations, roads, etc have startling differences in the ratio of injured people to total reported accidents or injuries to fatalities, or if those ratios change over time, some estimates of under reporting can be made. The same can be done for different regions of the United States with similar boating populations and on water activities AND ratios for the same region can be compared over time.
Issuing the Charge to Student Researchers
We are asking engineering, design, statistics, mathematics, and safety students to consider one or both of the approaches below as a Senior Design Project, Capstone Project, Senior Thesis, Masters Thesis, or similar project.
Project 1. Compare the results of two or more independent databases of propeller accidents, then use statistics to estimate the total number of accidents. Additional information on this approach is in the references near the bottom of this post. Independent databases could include:
- USCG BARD (for a period prior to USCG capturing newspaper reports and encouraging states to follow up on unreported accidents)
- Our (PGIC) database of media reports of propeller accidents
- NEISS (National Electronic Injury Surveillance System) – does not currently collect boat injury data
- ENA (Emergency Nurses Association) – collected boat and propeller injury data a few years ago.
- Several regional studies of propeller injuries
- Boat U.S. insurance database. See their comments at Beyond the BARD. February 12, 2008. USCG Accident Mitigation Meeting. (available from references below).
- Databases belonging to other major boat insurance companies
- Colossus – major insurance claims database used by insurance companies to estimate settlements based on past settlements for personal injury claims.
- There are probably others.
A comparison study of independent databases including USCG’s BARD might best be done by focusing on injuries and fatalities separately. With an accident more likely to be reported the more severe it is, fatalities should be reported.
A problem in comparing against BARD is they purposefully exclude several accidents (commercial boats, on private lands, boat involved in a crime, etc.). One would need to be careful you were not finding accidents in an outside database that had been reported to BARD, but excluded from Public BARD. Student researchers might need to work directly with USGC on those issues. We supply some details later on this post on the boat propeller accidents that are not listed as such in BARD.
With the databases supplying annual data, it would be easy to run multiple studies, one year at a time to gain an understanding of how underreporting has changed over time.
Project 2. Compare the ratios/percentages of propeller accidents, propeller injury frequencies, and propeller fatality frequencies across time or across regions. For example, if two states with relatively similar geography, populations, waters, etc have startling differences in the ratio of injured people to total reported accidents or injuries to fatalities, or if those ratios change over time, some estimates of under reporting can be made. Ratios for the same region can also be compared over time. These studies could be performed on BARD data or by analyzing some of the many existing studies.
If anybody is interested in making a Senior Design Project, Senior Thesis, Masters Thesis, Capstone of similar project from one of these two opportunities, please contact us using the contact tab in the top menu.
If neither of these “floats your boat”, but you are interesting in a boating safety project, please see several other similar projects listed in our Research Projects category.
Recreational Boat Propeller Accidents Not in BARD as Propeller Accidents
This post focuses on the accidents never reported to BARD, however, an understanding of those purposefully or accidentally excluded from BARD will be necessary to calculate some of the estimates.
Three streams of recreational boat propeller accidents are not reported as propeller accidents in BARD.
- Those not reported to BARD.
- Those reported to BARD but not correctly classified as a propeller accident in BARD
- Those reported to BARD that fail to meet USCG’s requirements to be listed in BARD
A boating accident report (BAR) is supposed to be filed for boating accidents that meet any of these criteria:
1. A person dies; or
2. A person disappears from the vessel under circumstances that indicate death or injury; or
3. A person is injured and requires medical treatment beyond first aid; or
4. Damage to vessels and other property totals $2,000 or more; or
5. There is a complete loss of any vessel.
As mentioned earlier there is broad disagreement about the percentage of recreational boat propeller accidents that are actually reported to BARD. The industry says near 100 percent, some propeller safety activists say “only a small percentage”.
We have previously identified several recreational boat propeller accidents in BARD that are in BARD, but not classified as propeller accidents in BARD. Our Misclassified Lewisville Lake Propeller Accident page describes on example. U.S. Coast Guard 2010 BARD Unreported Propeller Accidents: Carla Faul Pontoon Boat Propeller Accident in Florida is another.
Per USCG Recreational Boating Statistics 2010, if an accident is reported, it will be excluded from BARD if it meets any of the “Non-Reportable” criteria below:
- IRTAD Special Report. Underreporting of Road Traffic Casualties. Derriks and Mak of the Netherlands. International Traffic Safety Data and Analysis Group. International Transport Forum. June 2007. – compares independent databases
- Underreporting of Maritime Accidents to Vessel Accident Databases. Hassel, Asbjornslett, and Hole. Accident Analysis and Prevention. 2011. (Not yet published, and Elsevier Publication). – compares independent databases
- Underreporting of Maritime Accidents to Vessel Accident Databases. Martin Hassell. Masters Thesis. Norwegian University of Science and Technology. Department of Marine Technology. – basis of the article above.
- Underreporting of Maritime Accidents. Psarros, Skjong, Eide. Accident Analysis and Prevention. Vol.42. (2010) Pgs. 619-625. compares independent databases.
- UK Water-related Incident Database Gap Analysis. Final Report. RMS/06/14. Health & Safety Labor. Julie Bell, project leader. 2006.
- Investigating the Effects of Underreporting of Crash Data on Three Commonly Used Traffic Crash Severity Models: Multinomial Logit, Ordered Probit and Mixed Logit Models. Ye and Lord. Texas A&M University. Final Revised March 2011. – studies underreporting based on severity in automobile accidents
- Beyond the BARD. February 12, 2008. USCG Accident Mitigation Meeting. – explores some of the insurance databases, including Boat U.S..
- Recent Research on Recreational Boating Accidents and the Contribution of Boating Under the Influence. Summary of Results. Bruce Lawrence and Ted Miller. July 2006. A USCG funded product that examined BARD data.