Simulation Improves Estimation

Using a simulated population to evaluate survey design and analysis: a case study with Great Lakes fish.
Simulation is a useful tool for investigating the performance of a variety of statistical methods. A simulated population can provide valuable insight into the design and analysis of a survey, and may be employed in a wide range of applications. In addition, the process of defining a simulated population can bridge important gaps in understanding between the statisticians and technical experts involved. I present an example of a simulated fish population used to evaluate different methods of combining concomitant hydroacoustic and midwater trawl surveys. I worked with fisheries researchers to develop a fish community simulator that populates an artificial lake with known numbers of fish. The population was sampled by virtual hydroacoustic and midwater trawl surveys, and the methods used to combine the surveys were evaluated by comparing the resulting estimates to the known values.
This video was presented as an electronic poster at the American Statistical Association's Conference on Statistical Practice.