Is the signal real?
An increasing trend does not necessarily mean that the population is increasing.
Is removing noise enough?
Even when we remove as much of the noise as we can (i.e. by accounting for region and garden-type) it still appears from the adjacent Figure as though tūī are increasing slightly. But are they really?
There are methods available to us that tell us how likely it is that any observed increase is actually due to chance.
We can investigate this by determining what patterns we could likely see if tūī were not in fact increasing but remaining stable.
Essentially we are including a degree of uncertainty with our trend.
Measuring uncertainty?
The adjacent figure shows our best estimate of trend when we account for region and garden-type.
This time however we have included our uncertainty about this trend, as indicated by the grey shaded area.
Some of this grey shaded area is below the zero change line, indicating that there is a chance that tūī are in fact decreasing, despite our best estimate indicating a slight increase.
Our conclusion therefore is that we cannot be reliably sure as to whether tūī are increasing or decreasing.