At 12:35 AM 9/19/02 -0700, Juhani wrote: >Hi Everyone > >I need advice on how to verify a statistical experiment. > >The question is to determine whether emission reductions from two >different fuels are statistically significant. > >The test procedure was: the emissions of 12 similar engines were >tested twice, first with a normal fuel (Fuel A) and second time with >an improved formula (Fuel B). The results showed that on average the >reductions in emissions when using fuel B was significantly lower, >however I know enough statistics that this does not necessarily mean >the results conclusively demonstrate that Fuel B is better than Fuel >A. I would need to prove that the results is statistically valid > >Can anybody advice me in how to do this?
1. since using one fuel MAY impact on how the engine reacts to the OTHER fuel ... you should have have a splitting so that about 1/2 the tests had fuel A first then fuel B .. and the other 1/2 the reverse ... you want to make sure that this factor of timing is not important 2. with a modification like #1 ... do it again? do you get the same results? 3. you can NEVER EVER prove that one fuel is better than another but ... if several independently run experiments show the same results ... you're getting there >. >. >================================================================= >Instructions for joining and leaving this list, remarks about the >problem of INAPPROPRIATE MESSAGES, and archives are available at: >. http://jse.stat.ncsu.edu/ . >================================================================= Dennis Roberts, 208 Cedar Bldg., University Park PA 16802 <Emailto: [EMAIL PROTECTED]> WWW: http://roberts.ed.psu.edu/users/droberts/drober~1.htm AC 8148632401 . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
