Just to play devil's advocate, the only problem I have with that research is that they coach people on what constitutes good sound quality. I'm just wondering how different the results would be without instructing people on what to listen for.
In a true double blind test there would be no information given to participants nor there would be any interaction between the experimenter and participants. The goal is to create an unbiased testing environment as possible to ensure accurate results/minimize error/bias or any other anomalies.
