This study suggests an attention economy perspective on incomplete information, asserting that people use partial information as basis for action, hence an effective information system should provide them just with the most useful partial information, and should avoid redundant information that wastes their limited attention resources. These assertions were empirically examined via simulation of a speech-reading process that measured the real and perceived value of two levels of partial speech-reading support. The findings indicate that additional partial information improved performance, but there was practically no significant difference between one and two signals support levels. High correlation was found between the different methods of evaluation, suggesting that perceived value may be used as substitute to real value measurement after considering the circumstances of the system involved. This research may provide important insights for designing information systems in general, as well as speech-reading support systems that will improve communication opportunities of hearing impaired people.