To understand the ecological effects of forest restoration treatments on several old-growth forest stands in the Flathead National Forest of western Montana, USA, we surveyed birds at 72 points in treatment and control stands, and at more than 50 points in each of five potential reference stand conditions. We used a Before–After/Control-Impact design to assess treatment effects based on data collected 3 years before and 2 years after treatment. We also examined the similarity in bird community composition among all stand types by using a nonmetric multidimensional scaling approach. Relative abundances of only a few bird species changed significantly as a result of restoration treatments, and these changes were characterized largely by declines in the abundances of a few species associated with more mesic, dense-forest conditions, and not by increases in the abundances of species associated with more xeric, old-growth reference stand conditions. Thus, bird communities in treated stands were more similar to those in untreated stands of the same forest type than to those found in any of the potential old-growth reference stands. Although more time may be required for some bird species to respond to treatments, our results suggest that treatment plot sizes were either too small to affect bird communities or that the forest type selected for treatment was not within the range of forest types that are well suited for this type of forest restoration.