You have extremely large (or small) values that when transformed are driving the estimation of lambda.  On back-transforming, you may get an exponential overflow. Steve Denham ... View more

Yes, you're right in your interpretation. ... View more

Umm, I don't think they are all that different on the original scale.  Try this: lsmean scale01*year/ilink cl; This should up the display to include the mean on the original (untransformed) scale. Now, so far as selecting the proper distribution.  That is art, as much as science.  Plots of data, examination of residuals, consideration of the physical processes that generate the data and interpretability all enter in.  The usual things, like comparing various information criteria, are not so useful, as they depend on the form of the data, and once things are "transformed" as in a generalized linear model, it is like comparing apples to watermelons. One thing that might tell you how well things are fitting is the length of the confidence bounds, on the original scale.  Shorter means a more precise estimate--but again that is "art" and not rigorous, and tells you little about the accuracy of the estimation. And be sure, changing the distribution will have drastic effects on the location estimate. Steve Denham ... View more

You may be looking at special missing values. Look for example at the result of this test: data test; do x = 123, .a, .e, .i, .o, .u;     output;     end; run; proc print data=test; run; PG ... View more

The score type3 test is a test of the predictor's association with the response.  This is generally what you would use to assess the value of having the predictor in the model.  It is not a test of its effect on the correlation among the measurements.  But adding or removing predictors may effect the estimates of the correlation.  The QIC statistics don't have a measure of variability, so it is not possible to say if adding or removing a predictor has a "significant" effect on QIC.  ... View more

Dear all, I just read its SAS note. Does it mean I need to perform listwise deletion on my data before using the polychor macro? Thank you very much. ... View more

Think of the betas here as standardized beta hats in a regression, particularly if the data are rescaled and recentered, as is usually the case in a factor analysis.  This means that they are supposedly dimensionless, so that you can add things like wavenumber and polysaccharide concentration and not have to carry all of the units across all of the variables. Steve Denham ... View more

The answer always depends on the SAS release you are using.  Please include program and sample data so we can run the use case.  In the graph above, YEAR is a group role, so values for 2000 and 2001 are plotted as side by side clusters. They do not have specific axis values. Your best bet is to plot a graph using SGPANEL, with PANELBY=item01.  Then, each panel can have a SERIES x=year y=amount.  You can use other options to move the header values down to look like axis values, and suppress cell borders, etc. Potentially, if you have SAS9.40M1, you could define data for a POLYGON plot using discreteoffset=column to draw lines between each group value, but that may be more laborious, but better than SGAnnotate. ... View more