The scientific motivation, data collection strategy, data reduction, and analysis methods are presented for the Galactic Plane Infrared Polarization Survey (GPIPS). The chief goal for the Survey was to reveal the nature of the magnetic field threading the Galactic disk, in particular through regions of low to moderate extinction (1-20 mag of A V ) and star formation in the cool interstellar medium. The Survey region spans 76 deg 2 of the northern Milky Way disk, from = 18 • to 56 • and b = −1 • to +1 • . Linear polarimetric imaging observations began in 2006 in the near-infrared H band (1.6 μm) using the Mimir instrument on the 1.8 m Perkins telescope, located outside Flagstaff, AZ. Mimir used a cold, fixed wire grid and a rotateable cold, compound half-wave plate to obtain "stepand-integrate" polarimetry over its full 10 × 10 arcmin field of view. The GPIPS bright and faint polarimetric limits are approximately 7th and 15th mag, respectively, set by saturation and photon noise. Polarimetric uncertainties track with stellar magnitude, from about 0.1% to 25%, on average, from the brightest to faintest stars. Across the 3237 field GPIPS region, approximately 0.5 million stars are estimated to show detectable linear polarization (P /σ P > 3); most of these have m H < 12. This represents many orders of magnitude improvement in the number of polarization measurements across this region. GPIPS observations are more than 90% complete and should finish in 2012.

Methods and observations are described for the full field of view (FOV) polarimetric calibration of the Mimir nearinfrared imaging polarimeter in support of the Galactic Plane Infrared Polarization Survey and other applications. Polarimetric calibration consisted of three steps: (1) flat fielding using in-dome images obtained with the compound zero-order half-wave-plate (HWP) in the 16 position angles employed in polarimetric observations, (2) mapping and removing the remaining instrumental polarization via observations of globular cluster stars, and (3) using polarization standard stars to convert instrument-based polarization position angles to equatorial and to determine the linear polarimetric efficiency. For Mimir, the polarization flat fielding reduced systematic variations of the polarization efficiency across the FOV by a factor of 20. The 151 observations of eight globular clusters yielded 40,000 stars for measuring the remaining 0.05%-0.45% instrumental polarization to uncertainties of 0.02%-0.04%. After these corrections, the 444 observations of 23 polarized stars enabled equatorial polarization position angles to be determined to typical uncertainties of 0. • 5, and the on-axis (corrected to full FOV) polarization efficiency of 91.1% ± 0.4% to be determined. Fully corrected standard star observations showed excellent agreement with published values of polarization percentage and position angle. Additionally, the observations of four sky fields containing Whittet et al. "primary" polarization standard stars were analyzed to yield 30 new "secondary" standards. These are fainter than the primaries, allowing use with larger telescope apertures. The secondary standards have polarization position angle uncertainties under 5 • and range in degree of polarization from 0.4% to 8.5%.