I'm talking to investigators who measure the percent fibrosis of an organ
(between 0 and 100%). They want to compare the effect between treatment
groups; in particular, they would like to see a 20% or better difference.
We can't use logistic regression here, since the outcomes are not zero's or
one's. On the other hand, we cannot quite use linear regression either. I
know that if the sample size is large enough, we can (because of the central
limit theorem); but they're talking about a clinical trial of <100 patients.
a) What is the best method here?
b) How should one go about doing power calculations?
Re: a message to Jay ... A 1988 housing audit by the Urban Institute found that 59 percent of black home buyers and 56 percent of Hispanic home buyers encountered discrimination when searching for a home. ... When the treatment of these individuals was compared, it was found that blacks and Hispanics were more likely than whites to be told no housing was available in their price range. ... Do you think you can even begin to compare the treatment of slaves and ... (uk.sport.football.clubs.liverpool)
Re: a message to Jay ... A 1988 housing audit by the Urban Institute found that 59 percent of black home buyers and 56 percent of Hispanic home buyers encountered discrimination when searching for a home. ... When the treatment of these individuals was compared, it was found that blacks and Hispanics were more likely than whites to be told no housing was available in their price range. ... Do you think you can even begin to compare the treatment of slaves and ... (uk.sport.football.clubs.liverpool)
Re: Effect size with no control group? ... Because the variation in the column of differences was ... if you compare "After" against a standard ... comparisons are made of "no treatment vs a treatment" to demonstrate ... suppose that a control group is strictly needed ... (sci.stat.consult)
Re: Effect size with no control group? ... There are two distinct notions of "the effect of treatment."... And it is bad experimental design,... suppose that a control group is strictly needed ... Test 1 would act as "control group" to compare with Test 2. ... (sci.stat.consult)
John Chewter and Whitening software ..... ... If the treatment is finished ... take a second photo of the result situation - to compare pre- and ... Further with a special Software you can simulate a possible result of ... to sell the treatment and show & compare the ... (sci.med.dentistry)
