We present an analysis of the linear polarization of six active galactic nuclei -0415+379 (3C 111), 0507+179, 0528+134 (OG+134), 0954+658, 1418+546 (OQ+530), and 1637+574 (OS+562). Our targets were monitored from 2007 to 2011 in the observatory-frame frequency range 80-253 GHz, corresponding to a rest-frame frequency range 88-705 GHz. We find average degrees of polarization m L ≈ 2−7%; this indicates that the polarization signals are effectively averaged out by the emitter geometries. From a comparison of the fluctuation rates in flux and degree of polarization we conclude that the spatial scales relevant for polarized emission are of the same order of, but probably not smaller than, the spatial scales relevant for the emission of the total flux. We see indication for fairly strong shocks and/or complex, variable emission region geometries in our sources, with compression factors < ∼ 0.9 and/or changes in viewing angles by > ∼ 10• . An analysis of correlations between source fluxes and polarization parameter points out special cases: the presence of (at least) two distinct emission regions with different levels of polarization (for 0415+379) as well as emission from a single, predominant component (for 0507+179 and 1418+546). Regarding the evolution of flux and polarization, we find good agreement between observations and the signal predicted by "oblique shock in jet" scenarios in one source (1418+546). We attempt to derive rotation measures for all sources, leading to actual measurements for two AGN and upper limits for three sources.We derive values of RM = (−39 ± 1 stat ± 13 sys ) × 10 3 rad m −2 and RM = (42 ± 1 stat ± 11 sys ) × 10 4 rad m −2 for 1418+546 and 1637+574, respectively; these are the highest values reported to date for AGN. These values indicate magnetic field strengths of the order ∼10 −4 G. For 0415+379, 0507+179, and 0954+658 we derive upper limits |RM| < 1.7 × 10 4 rad m −2 . From the relation |RM| ∝ ν a we find a = 1.9 ± 0.3 for 1418+546, in good agreement with a = 2 as expected for a spherical or conical outflow.