It's easiest to understand the model, I think, in two steps, although it is estmated in one.
Y(i,j) = b(0,i) = b(1,i)*t(j) + u(i,j) is the level one model, assuming linearity over time. B(0,i) = G(0,0) + G(0,1)*trt + e(0,i) B(1,i) = G(0,1) + g(1,1)*trt + e(1,i) Where the bs are random growth parameters, t(j) is time j, u is a random error at level 1, G(0,k) is a fixed component of the random variable, and trt is a treatment indicator. E is a n error term at level 2. This model should be estimable in SPSS version 11 but since I don't use SPSS, I do not know the syntax. If the nature of change is not lllinear, then a different level 1 model must be specified. Paul R. Swank, Ph.D. Professor, Developmental Pediatrics Medical School UT Health Science Center at Houston -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Frank Rigous Sent: Tuesday, December 09, 2003 4:05 AM To: [EMAIL PROTECTED] Subject: Re: Classical design? [EMAIL PROTECTED] (Paul R Swank) wrote in message news:<[EMAIL PROTECTED]>... Thanks! > What is the nature of the change over time? The nature is a) the treatment (hopefully) and b) the initial change (decrease or increase) after onset of the disease. > If it is linear then a mixed > models analysis looking at change modeled individually for each > repsondent would work. This way, the time between assessments does not > have to be the same across assessments and missing (at random) data > can be handled. This could be done with SAS, SPSS (11), Mlwin, or HLM5 > at the least. You mean, calculating the solpes for every subject and the interaction between the two groups? Or are the individually slopes nested under the treatment? Is this the following model? y(ij) = t(i) + x(ij)*b(ij) + e(ij) y(ij): response to timepoint i=2,3,4, subject j t(i): treatment effect x(ij)*b(ij): time x(ij) times slopes b(ij) nested under t May you give me a sample syntax in SPSS for your suggestion? (Problems are most often in details.) Frank Rigous > > Paul R. Swank, Ph.D. > Professor, Developmental Pediatrics > Medical School > UT Health Science Center at Houston > . . ================================================================= 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/ . =================================================================
