Table 1 Examples of how to address uncertainty for different datasets
Data type | Available data | Possible estimation of uncertainty and variability | Source of uncertainty which can be addressed |
|---|---|---|---|
Litter size (common vole, M. arvalis) | 1 to 2 (N = 37) 3 to 4 (N = 106) 5 to 6 (N = 44) 7 to 8 (N = 7) [68] | Confidence intervals can be calculated after recreation of the original sample underlying the data, and variability can be estimated based on the data distribution (or calculation of a standard deviation) | Only uncertainty due to sample size is included; uncertainty due to study location, climate or other factors is not included |
Clutch size (little owl, Athene noctua) | Central Europe, usually 3 to 5 and exceptionally up to 7 [69] France, 3.9 (N = 80, [70]) Portugal, 3.3 ± 1.2 (min, 1; max, 5; N = 15, [71]) | Only qualitative analysis of uncertainty is possible. However, the standard deviation from the study of Tomé et al. [71] and the ranges from Glutz von Blotzheim [69] indicate that an average clutch size between 3 and 4 is realistic. Data from Portugal indicate that smaller clutch sizes might be observed in southern Europe | Variability and uncertainty due to sample size cannot be distinguished in detail, but sample sizes and comparable ranges indicate that data are reliable. Since data from several studies and years are shown, uncertainty due to study location and temporal variability can be estimated at least qualitatively |