This study compared two methods of providing analysts with ATR (Automatic Target Recognition) confidence ratings, and examined whether these ratings yielded superior performance compared to ATR designations provided without added confidence ratings. Twelve expert analysts were presented with SAR (Synthetic Aperture Radar) images, each of which contained one of three types of ATR designations. Two of the designation types included ATR confidence ratings and a third did not. The two ATR designation types that included confidence levels specified the confidence levels as either below.70, about .80, or over .90. The confidence level was indicated either by surrounding the SAR item with three different shapes or by marking the item with a number. The third designation type did not include information about confidence levels, but rather used an ellipse to matk all SAR items designated by the ATR as targets. None of the perfromance measures of performance, i.e., Hit Rates (HR), False Alarm Rates (FAR) or the signal-detection statistic d' yielded major differences between the three designation types, with the exception of slightly (but borderline significantly) fewer FARs for the ellipses. To test whether ATR confidence ratings were not heeded by the participating analysts, all instances in which designated targets included ATR confidence levels were combined (numbers and shapes), and HRs and FARs were calculated for each of the three confidence levels. Both HRs and FARs decreased as confidence levels decreased, as had been shown in our earlier study (Setter, Norman, & Marciano, 2004). This indicated that the analysts did indeed acknowledge the ATR confidence designations. A claim was made that although the results of this study seem to indicate that ATR confidence designations do not contribute to analysts' performance, there are valid reasons to assume that they would improve performance in a "real life" situation.