A Quality Measurement Experiment (QME) comparing clear-sky downwelling longwave radiance at the surface from observations and model calculations has been performed by the Atmospheric Radiation Measurement (ARM) Program for many years (e.g., Brown et al. 1995; Clough et al. 1996). This QME has been used to (1) validate and improve absorption models and spectral line parameters used within the Line-by-Line Radiative Transfer Model (LBLRTM), (2) assess the ability to define the atmospheric state, and (3) assess the quality of the atmospheric emitted radiance interferometer (AERI) observations. Extensive analysis of a 1994 to 1997 dataset has highlighted many issues associated with the ability to specify the atmospheric state (primarily water vapor) in the model and to characterize the influence of thin clouds and aerosols. The QME also revealed some uncertainties in the AERI observations themselves. Many of these issues have been addressed via reprocessing or various correction techniques. However, the uncertainties that remain in this dataset prohibit the use of the QME for its original goal: the improvement of the radiative transfer model itself, particularly with respect to the longwave self-broadened water vapor continuum absorption. This work concentrates on the analysis of a new QME dataset, which addresses some of the uncertainties in the older dataset. The new QME dataset consists of 241 nighttime only cases from 1998 to 2001 from the Southern Great Plains (SGP) Central Facility. Coincident and quality controlled AERI, Vaisala radiosonde, ARM Raman lidar, and microwave radiometer (MWR) observations exist for each case. The dataset is restricted to nighttime only because daytime water vapor profiles from the Vaisala radiosondes have higher variability and larger biases with respect to the MWR, as inferred from long-term comparisons of radiosonde and MWR precipitable water vapor (PWV) measurements. The Raman lidar water vapor profiles also extend to higher altitudes (~12 km) at night, but only to ~3.5 km during the day. The depolarization sensitivity of the Raman lidar is used to screen for thin clouds. The Raman lidar also provides aerosol optical depths for each case, allowing the effects of aerosols on the infrared residuals to be characterized. MWR tip curve data and variability of AERI data over the averaging period is used to screen for sky inhomogeneity. In contrast to the 1994 to 1997 dataset, this time period also includes MWR processing, including consistent tipping-curve operation and processing and application of beam width and other small corrections (Liljegren 1999).