NESDIS and CIMSS have produced GOES imager clear-sky brightness temperatures (CSBT) in WMO BUFR format since November, 2001. The data include 1 visible and 4 infrared channels. These data were examined for possible use in the NCEP global data assimilation system (GDAS). Larger negative differences have been found with larger cloud fraction in the data when comparing the data with CSBT simulated from GDAS fields. The characteristics of the differences suggest cloud contamination in the data. Data quality control procedures were developed after studying observed CSBT and the differences between observed and simulated CSBT. For the infrared moisture channel, channel 3, 15% of the data were eliminated because of what is believed to be cloud contamination. For the infrared channels 2, 4 and 5, significant cloud contamination remains even after substantial quality control. Thus, only data from channel 3 are currently evaluated for use in the GDAS. The feedback of these results and others has prompted NESDIS and CIMSS to reexamine cloud clearing algorithm. Data assimilation cycling and forecast experiments evaluating the impact of these data within the NCEP GDAS are underway and will be presented.