Static and dynamic parameters of compression can interact in complex ways to affect the sound quality of speech and other signals, especially in background noise. The complexity of these interactions presumably increases with number of compression channels; consequently, there currently is a lack of general agreement on how to select parameters of multichannel compression to optimize speech quality in noise. Results from two major research projects will be reported: 1) a survey of 2,000 dispensing audiologists across North America showed conclusively that there is fundamentally no agreement among practitioners on how to set compression parameters in hearing aids, or how to interpret users' complaints regarding sound quality in compression hearing aids; 2) a large-scale laboratory study of listeners' preferences for speech in noise processed via two-channel compression showed that several combinations of release times and compression ratios could produce good-quality processed sound, just as several other combinations were reliably judged by users as objectionable. A quantitative model - based on the preference data - was constructed for linking the release times and compression ratios of a two-channel compression system to the preference judgments of normal-hearing and hearing-impaired listeners. The model can potentially be used to select successful combinations of compression parameters based on intelligent combination of release time and compression ratio, rather than separate selection of each.