Knowledge about the anisotropy of VHF radio wave scatterers in the atmosphere is important when VHF wind profiler radars are used to measure atmospheric winds by Doppler beam-swinging methods, since appropriate correction factors must be employed. In this study, the anisotropy of the scatterers is determined using two different methods, over the course of a full year, for a VHF radar at Resolute Bay in northern Canada (75 ¿N). The first method utilizes direct comparisons of the radar-derived winds with those of radiosondes launched from close to the site, while the second uses comparisons of powers received on vertical and off-vertical beams. The study is unique in that a full year of radiosonde data were available from balloons which were launched from a site only 4 km from the radar. The two methods for estimating the anisotropy at first appear to give slightly different conclusions, but by properly considering the errors associated with both radar and radiosonde measurements it is possible not only to reconcile the two sets of measurements but also to give reasonable estimates of the errors associated with both the sonde and radar winds. An important result which arises is that the radar data have very similar accuracies to the radiosonde data. Furthermore, it is demonstrated that in the past, radar wind measurements may have been overcorrected, because the errors in the sonde measurements may not have been properly considered in determining the correct conversions. Seasonal variations in the aspect sensitivity are found but are not large. ¿ 2001 American Geophysical Union