Interpreting differences between the maps from the first breeding atlas in the 1980s and those from the second breeding atlas is fraught with difficulty because we lack information on survey effort during the first atlas. Survey effort involves not only the amount of time spent in the field, but also how and where that time was spent. As a result, direct estimation of trends from atlas data is very hard. When there have been very large increases (e.g., Osprey, Pine Warbler) or decreases (e.g., Ruffed Grouse, Canada Warbler) in the range of a species, then it is reasonable to assume that the change is real. More subtle differences in the number of occupied blocks, however, are much more difficult to interpret. For this reason, we do not report percentage changes in the number of occupied blocks for all species.

Changes in the level of breeding evidence – for instance the number of blocks with confirmed breeding – are especially problematic because they are influenced both by the amount of effort and when during the breeding season the surveys took place. In general, we found that breeding confirmation rates were lower in the current atlas than in the first atlas. Given that, for most species, there is no reason to expect species to forego breeding when they are present in an area, this result likely reflects a difference in observer effort, rather than any real change in breeding behavior.

In the species accounts, we have provided our best assessment of what observed changes really mean, generally by looking at the atlas block data in combination with other sources of information. We also considered how likely it is that an observed change reflects a real pattern, based on characteristics of the species. For instance, we assumed that observed changes are more likely to be reliable indicators of real patterns for species that are conspicuous and easy to identify than for those that are not, since detection of these species is less likely to be affected by the amount of search effort.

We also took note of broader patterns in the data when interpreting apparent differences between the two data sets. For instance, during the first atlas, blocks in the northwest corner of the state consistently had many species reported than blocks elsewhere. We do not know the reasons for this, but the pattern is sufficiently pronounced  that it seems likely that search effort was especially high in that region. The northeast and east-central portions of the state were also reported to have received less coverage than other parts of the state (Bevier 1994). In contrast, although coverage varied among blocks in the current atlas, differences tended to vary more with human population density – with coverage generally greatest towards the coast, and in the center of the state. Consequently, within-state shifts in relative abundance of species should be viewed through the lens of these likely variations in within-atlas survey effort.