Botta-Dukát Z, Kovács-Láng E, Rédei T et al: Statistical and biological consequences ... (2007)

Botta-Dukát Z, Kovács-Láng E, Rédei T, Kertész M, Garadnai J
Statistical and biological consequences of preferential sampling in phytosociology: theoretical considerations and a case study
Folia Geobotanica 42: 141-152

Due to the long tradition of the Braun-Blanquet approach, many relevés using this approach have been made. Recent developments in vegetation-plot databases provide an opportunity to effectively use these relevés to study ecological problems as well. Opinions differ, however, concerning the applicability of these datasets, often with their use being restricted to exploration and hypothesis generation only.
We assert that preferential sampling, which is characteristic of the Braun-Blanquet approach, means using a special definition of statistical population rather than non-random sampling. We present a case study, where consequences of using a preferential and non-preferential definition of statistical population are studied.
Although the traits of stands that are preferred or avoided by the phytosociologist during preferential sampling can be identified, there are no general rules that could predict the difference between the preferential and non-preferential datasets obtained for the same object.