We compare geodetic and single-station methods of measuring dynamic deformations and characterize their causes in the frequency bands 0.5--1.0 Hz and 4.0--8.0 Hz. The geodetic approach utilizes data from small-aperture seismic arrays, applying techniques from geodesy. It requires relatively few assumptions and a priori information. The single-station method uses ground velocities recorded at isolated or single stations and assumes all the deformation is due to plane-wave propagation. It also requires knowledge of the azimuth and horizontal velocity of waves arriving at the recording station. Data employed come from a small-aperture, dense seismic array deployed in Geyokcha, Turkmenistan, and include seismograms recorded by broadband STS2 and short-period L28 sensors. Poor agreement between geodetic and single-station estimates in the 4.0--8.0 Hz passband indicates that the displacement field may vary nonlinearly with distance over distances of ~~50 m. STS2 geodetic estimates provide a robust standard in the 0.5--1.0 Hz passband because they appear to be computationally stable and require fewer assumptions than single-station estimates. The agreement between STS2 geodetic estimates and single-station L28 estimates is surprisingly good for the S-wave and early surface waves, suggesting that the single-station analysis should be useful with commonly available data. These results indicate that, in the 0.5 to 1.0 Hz passband, the primary source of dynamic deformation is plane-wave propagation along great-circle source-receiver paths. For later arriving energy, the effects of scattering become important. The local structure beneath the array exerts a strong control on the geometry of the dynamic deformation, implying that it may be difficult to infer source characteristics of modern or paleoearthquakes from indicators of dynamic deformations. However, strong site control also suggests that the dynamic deformations may be predictable, which would be useful for engineering seismically resistant structures.