Making the best of bad data: A longitudinal analysis of elephant poaching in Africa from 2002 to 2018

Abstract

The issue of bad data and the potential misinterpretation of that data is becoming part of a growing debate within the wildlife conservation field. However, research on this topic has been limited compared to other fields. Therefore, to overcome these issues, this study followed an inductive data-driven approach using mixed-effects longitudinal methods for count-based data. This study adds to this growing debate by attempting to discern the potential causes of elephant poaching in African countries utilizing potentially unreliable data. Because wildlife crime is a societal problem, this study measured the probability of a wildlife reserve reporting illegally killed elephants found on MIKE (monitoring of illegally killed elephants) sites, given country-level factors—such as the amount of gross domestic product (GDP), the rate of urbanization, the total percentage of forest area in a country, and the perception of corruption in the country—and wildlife reserve-level factors, such as the size of the MIKE sites and the population estimates for the number of elephants within these MIKE sites. This study also looked at other controlling factors, such as misreporting information, the reliability of reporting information, and the international effect on elephant poaching. As a result, this study found that several factors attributing to economic demand, urbanization, corruption in the country, and the elephant population size on a designated MIKE site contributed to the odds of elephants being reported as having been illegally killed despite issues with reporting, international pressure, and efforts to crack down on corruption in end-market countries like China.