The following files are provided in PKZIP compressed file formats which can be saved to your system by asking the web browser to download them. They require PKUNZIP to uncompress them.

MIXOR.ZIP has a batch version of MIXOR (version 2.0), a program to analyze multilevel data with a dichotomous or ordinal response. Some documentation is provided in this ZIP file (see files README.TXT and MIXORE2.PDF after unzipping). The program's executable is dated 12/9/98.

MIXREG.ZIP has a batch version of MIXREG, which assumes a normal theory model, but can deal with AR, MA, ARMA, and Toeplitz dependencies of the disturbances. Some documentation is provided in this ZIP file (see file README.TXT after unzipping). The program's executable is dated 11/19/98.

MIXGSUR.ZIP has a batch version of MIXGSUR, a program to analyze multilevel grouped-time survival data. The program's executable is dated 9/6/96. Note, all of the features of MIXGSUR are now included in MIXOR.

MIXNO.ZIP has a batch version of MIXNO—a program to perform mixed-effects nominal logistic regression analysis. Some documentation is provided in this ZIP file (see file README.TXT after unzipping). The program's executable is dated 1/4/99.

MIXPREG.ZIP has a batch version of MIXPREG—a program to perform mixed-effects Poisson regression analysis. Some documentation is provided in this ZIP file (see file README.TXT after unzipping). The program's executable is dated 8/20/98.

Manuals are posted here as PDF files.

There are also a few macro-type programs written by Don to fit random-intercepts models via SPSS and/or SAS. MIXREGPR.ZIP contains the following:

• MIXREG.SPS which uses SPSS Matrix routines to fit a random-intercepts regression model
• MIXREG.SAS uses SAS IML routines to fit a random-intercepts regression model
• MIXOR.SPS uses SPSS Matrix routines to fit a random-intercepts ordinal probit or logistic regression (unfortunately, this last macro program is very slow)

There is also a SPSS macro program to estimate proportional odds or partial proportional odds models (without random effects): THRESHM1.SPS

and a SPSS macro program to estimate non-proportional odds models (without random effects): THRESHM2.SPS

Also, a SAS macro program to estimate proportional odds or partial proportional odds models (without random effects): THRESHM1.SAS

and a SAS macro program to estimate non-proportional odds models (without random effects): THRESHM2.SAS

As applied to stages of change data, these models are described in Hedeker, Mermelstein, Weeks (1999) and Hedeker & Mermelstein (1998). See References 26 & 27

These macro-programs are pretty well self-documented.