We expand on the dispersion analysis of polarimetry maps toward applications to interferometry data. We show how the filtering of low-spatial frequencies can be accounted for within the idealized Gaussian turbulence model, initially introduced for single-dish data analysis, to recover reliable estimates for correlation lengths of magnetized turbulence, as well as magnetic field strengths (plane-of-the-sky component) using the Davis-Chandrasekhar-Fermi method. We apply our updated technique to TADPOL/CARMA data obtained on W3(OH), W3 Main, and DR21(OH). For W3(OH) our analysis yields a turbulence correlation length $deltasimeq19$ mpc, a ratio of turbulent-to-total magnetic energy $leftlangle B_{mathrm{t}}^{2}rightrangle /leftlangle B^{2}rightrangle simeq0.58$, and a magnetic field strength $B_{0}sim1.1:mathrm{mG}$; for W3 Main $deltasimeq22$ mpc, $leftlangle B_{mathrm{t}}^{2}rightrangle /leftlangle B^{2}rightrangle simeq0.74$, and $B_{0}sim0.7:mathrm{mG}$; while for DR21(OH) $deltasimeq12$ mpc, $leftlangle B_{mathrm{t}}^{2}rightrangle /leftlangle B^{2}rightrangle simeq0.70$, and $B_{0}sim1.2:mathrm{mG}$.