We present an analysis of the degree to which seabird species respond to ships, thereby affecting the results of at-sea strip transect surveys (continuous, snapshot) undertaken to quantify community patterns of tropical seabirds. Two survey methods for counting seabirds were employed simultaneously and independently on research cruises in the eastern tropical Pacific Ocean: (1) a 600-m strip quadrant was surveyed on one side of the bow using 8×40 binoculars (strip transect), and (2) birds were surveyed from the bow to the horizon on both sides of the ship using 25×150 mounted binoculars (“big-eyes”). Data collected using each method were compared to determine potential biases of the strip transect method, with particular attention paid to seabird ship avoidance. Seabird species were assigned to six categories to control for detection biases resulting from differences in body size and flight behavior, and comparisons between methods were made within each of these categories. Our results indicate that frigatebirds Fregata spp., Sooty Terns Sterna fuscata and White Terns Gygis alba may avoid ships, and therefore, unless compensatory procedures are taken, estimates of density and abundance of these species using strip transects may be negatively biased.