This dataset is a collection of data from 4-year experiments conducted in the Gironde department in southwestern France. It contains “Block, Year, Soil type, neighboring, Species and Origin” as explanatory categorical variables and “Target survival” as binomial response variable.

These data were analyzed in order to characterize and assess how the effect of species stress tolerance and community competitive effects lead to differences in species composition between calcareous and siliceous plant communities.

The year, rock and origin variables and their interactions as fixed factors, were used to assess the effect of abiotic conditions of the soil type and year quality, on the survival of target species from calcareous and siliceous soils growing without neighbors, using a general linear mixed-effects model with a logit link function for binomial distribution.

The survival with and without neighbours was also used to calculate two indices to characterize the intensity (RII) and the importance (Iimp) of plant-plant interactions. RII and Iimp were then analyzed separately using linear mixed-effects models with “Year quality”, “Soil type”, “Origin” and their interactions as fixed factors and “Year nested within Blocks” as a random factor.

All data were stored using Microsoft Excel and all the statistical analyses were done using R software (R Core Team 2016).