Don't just survey the easy spots: A seminar on why "random" matters

In environmental surveys, we rarely have the luxury of a "perfect" setup. In an ideal world, we’d select our study sites using strict probability sampling to ensure every location has a fair chance at being included. This keeps our results unbiased and gives us a solid scientific basis for our conclusions. In the real world, however, we often have to be pragmatic. We survey what’s easy to reach, what experts recommend, or the areas we know are rich in species. While these non-random methods save time and money, they come with a hidden cost: they can accidentally steer our results in the wrong direction. In this seminar, we explore what happens to our findings when our study sites aren't chosen at random. Using simulations based on a typical biodiversity monitoring setup, we show how different ways of selecting sites can either strengthen or limit our ability to make broad, population-level conclusions.

The seminar will cover:

• Definitions and comparisons of probability and nonprobability sampling.
• Statistical concepts such as bias, representativeness, and MSE.
• Simulations to illustrate the concepts in the setting of an ecological on-site survey.
• Implications of nonprobability sampling on ecological research and management.

Speakers: Peter Lundquist and Annica de Groote

Date and time: 2026-09-17 10:00-12:00 and 13:00-15:00.

Location: A hybrid meeting at Studio Ultuna and Zoom.

Registration is required to participate in the seminar.