Harley, Experimental psychologists always face the problem that the experimental situation is an approximation of the real world. Faced with the choice between (1) a good emulation of reality that almost certainly will not produce an interpretable experimental result and (2) a poorer emulation of reality that has a good shot at getting an interpretable result, the experimentalists I know will almost always take door number 2. I say it's an unstated de facto rule of thumb. Greg >Greg, > >You have hit the proverbial nail right on the head! The proponents in >favor of single-stimulus argue (and this is valid) that single stimulus is >more appropriate to the nature of RR/NR schemes where a comparison channel >is not available. Evidently the (just as valid) counter argument is that >not enough data can be collected in order to do what is required, i.e., >discriminate amongst HRC's. The real question is: how can both points of >view be accommodated realistically? In other words, given that you accept >the fact that the best way to maintain integrity of testing is to perform >the testing as it is used, how should the test be constructed so as to >give the maximum discrimination possible between HRCs. Are there no rules >of thumb to aid in this process? > >--Harley > >At 2:56 PM -0400 7/16/01, Greg Cermak wrote: >> >Dear Colleagues, >> > >> >I am in the process of drafting a test plan for the RR/NR-MM activity and >> >need a bit of feedback. I am patterning the plan after Jamal's excellent >> >RR/NR-TV plan and the document "Results of an audiovisual desktop video >> >teleconferencing subjective experiment" (vqeg008.doc/vqeg008.pdf). A big >> >questions is SSCQE or DSCQS (or some other method). This may be premature >> >as we have yet to determine the specific application areas. Some may argue >> >that the subjective testing should be the same as that of the RR/NR-TV and >> >that is fine, but i would like to get the dialogue initiated before the >> >actual meeting if possible. >> > >> >Thank you, >> > >> >-- >> >Harley R. Myler >> >Professor of Electrical Engineering >> >School of EE &CS >> >College of Engineering & CS >> >University of Central Florida >> >Orlando, Florida 32816-2450 >> > >> >>Harley and VQEG, >> >>From the point of view of data analysis, the NOT-continuous methods will be >>a lot easier to analyze with familiar methods. From the point of view of >>being able to discriminate among HRC's, Comparative (two-stimulus) methods >>should lead to better results than single-stimulus methods. >> >>Greg >>Gregory W Cermak >>Principal Member of Technical Staff >>Verizon Laboratories >>Mailcode LA0MS38 >>40 Sylvan Rd >>Waltham, MA 02451 >>(781) 466-4132 >>FAX (781) 466-4035 > > >_______________________________________________ >ituvidq mailing list >ituvidq@its.bldrdoc.gov >http://www.its.bldrdoc.gov/mailman/listinfo/ituvidq