Timing and location of reproduction in African waterfowl: an overview of >100 years of nest records.

In other words, to accept the hypothesis that species with more extensive ranges also show significantly greater variability in the timing of breeding, we would need to find (when comparing between different species): (1) a significant, nonzero, increasing trend between mean temporal distance and mean geographic distance; (2) the lack of a trend, or at least a significantly lower slope, for the null data set; and/or (3) an increasing trend in the standard deviations of the temporal distances for each species, with either the magnitude of the standard deviation increasing with increasing range extent or one or more species showing nonoverlapping standard deviations.

also in 63,971 other papers

To correct for possible sampling bias we also ran this analysis on a randomized “null” data set, with the expectation that if there were a significant trend in the actual data, the results would differ significantly from the null data.

also in 9,775 other papers