Datasets usually provide raw data for analysis. This raw data often comes in spreadsheet form, but can be any collection of data, on which analysis can be performed.
Death or injury to whales from vessel strike is one of the primary threats to whale populations worldwide. However, quantifying the rate of occurrence of these collisions is difficult because many incidents are not detected (particularly from large vessels) and therefore go unreported. Furthermore, varying reporting biases occur related to species identification, spatial coverage of reports and type of vessels involved. The International Whaling Commission (IWC) has compiled a database of the worldwide occurrence of vessel strikes to cetaceans, within which Australia constitutes ~7% (35 reports) of the reported worldwide (~471 reports) vessel strike records involving large whales. Worldwide records consist largely of modern reports within the last two decades and historical evaluation of ship strike reports has mainly focused on the Northern Hemisphere. To address this we conducted a search of historical national and international print media archive databases to discover reports of vessel strikes globally, although with a focus on Australian waters. A significant number of previously unrecorded reports of vessel strikes were found for both Australia (76) and worldwide (140), resulting in a revised estimate of ~15% of global vessel strikes occurring in Australian waters. This detailed collation and analysis of vessel strike data in an Australian context has contributed to our knowledge of the worldwide occurrence of vessel strikes and challenges the notion that vessel strikes were historically rare in Australia relative to the rest of the world. The work highlights the need to examine historical records to provide context around current anthropogenic threats to marine fauna and demonstrates the importance of formalized reporting structures for effective collation of vessel strike reports. This paper examines the issues and biases in analysis of vessel strike data in general that would apply to any jurisdiction. Using the Australian data as an example we look at what information can be inferred from historical data and the dangers of inference without consideration of the reporting biases.