Population age structure and vital statistics are important for understanding songbird demography and for developing conservation strategies. Field‐based estimates of survival rates based on mark–recapture methods are conservative because they are constrained by problems associated with detection probabilities and emigration. However, data collected at bird‐banding stations during spring and fall migration can potentially provide useful demographic information. I used banding data collected over a 6‐yr period (2005–2010) at Long Point Bird Observatory on the north shore of Lake Erie in Ontario, Canada, and Powdermill Avian Research Center in southeastern Pennsylvania, U.S.A., to determine if banding records could be used to estimate vital statistics for several species of songbirds. As reported in previous studies, I found the proportion of juveniles captured during fall migration to be unrealistically high to be representative of true proportions, especially at Long Point. The proportion of juvenile songbirds captured remained implausibly high during spring migration, with related estimates of longevity and generation time implausibly low and of fecundity implausibly high. Based on apparent adult survival estimates from the literature that suggest an average age structure for songbirds of >55% adults and <45% juveniles, I found that capture rates for juveniles during spring migration were at least twice as high as that for adults. A slower pace of spring migration by juveniles likely accounts for some of this bias. Because the data cannot be assumed to represent unbiased samples with respect to the age structure of populations, my results indicate that banding data collected at bird‐banding stations during migration are not suitable for use in demographic studies.