Yes, Kjetil, As you suggest at the end, the researcher can create a Regression/Linear Model, the Assumed model, so that the questions of interest can be expressed in terms of the parameters of the Assumed Model, then impose the desired restrictions to obtain a Restricted Model and then let the least squares program do the dirty work. SAS makes this easy, but any Regression Program can be used-- as long as it obtains the correct answers.
-- Joe ----- Original Message ----- From: <[EMAIL PROTECTED]> To: "King" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> Sent: Tuesday, November 04, 2003 4:31 PM Subject: Re: Help needed to analyze 2X6 factorial data! > On 3 Nov 2003 at 19:21, King wrote: > > For a book: Box, Hunter & Hunter: "Statistics for experimenters" > > This will not give you a cook-book guide for your unbalanced design, > but general principles. > > Start out with graphical analysis, compute the means for each of the > cells and use those for interaction plots (for each levbel of factor > A, draw the means for tle levels of factor B united with a line). The > same with roles reversed. > > A linear-model calculation for an unbalanced design by hand is no > fun, you should consider a simple least-squares program. > > Kjetil Halvorsen > > > > Help please!! I desperately need to analyze data from a 2X6 factorial > > design experiment (Factor A has 2 levels and Factor B has 6 levels). > > The experiment is replicated 5 times, and there is a variable number > > (between none to 3) data points per treatment combination per > > replicate. Are there any on line resources or text books with examples > > of worked-out problems available? (I'm required to do most of the > > calculations by hand). It would be better if the resources are in very > > simple terms, basic enough for a non-statistician like myself to > > comprehend. Any help with this would be greatly appreciated. Thanks!! > > . > > . > > ================================================================= > > Instructions for joining and leaving this list, remarks about the > > problem of INAPPROPRIATE MESSAGES, and archives are available at: > > . http://jse.stat.ncsu.edu/ . > > ================================================================= > > > . > . > ================================================================= > Instructions for joining and leaving this list, remarks about the > problem of INAPPROPRIATE MESSAGES, and archives are available at: > . http://jse.stat.ncsu.edu/ . > ================================================================= . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
