Weight Loss Data Set

Weight Loss data set 1:

• SAS Programs and Outputs
  • Make the sas data set.
    • SAS code to creat a SAS data set
    • Parallel plot of the data. (postscript)
  • Model 1 (Fixed Effect = Intercept, Week ; Random Effect = Intercept)
    • SAS program to fit the model
    • Unedited SAS output & Residual plot (postscript)
  • Model 2 (Fixed Effect = Intercept, Week WeekWeek ; Random Effect = Intercept)
    • SAS program to fit the model
    • Unedited SAS output & Residual plot (postscript)
  • Model 3 (Fixed Effect = Intercept, Week, WeekWeek, WeekWeekWeek ; Random Effect = Intercept)
    • SAS program to fit the model
    • Unedited SAS output & Residual plot (postscript)
  • Model 4 (Fixed Effect = Intercept, Week ; Random Effect = Intercept, Week )
    • SAS program to fit the model
    • Unedited SAS output & Residual plot (postscript)
  • Model 5 (Fixed Effect = Intercept, Week WeekWeek ; Random Effect = Intercept, Week )
    • SAS program to fit the model
    • Unedited SAS output & Residual plot (postscript)

Weight Loss data set 2:

Models 1-4 use ML fitting. Models 5-8 are the exact same models, but use REML fitting. Models 1,2 and 4 have the intercept and day random effects correlated with each other, but not with visit. Model 3 has all three random effects correlated.

Plots below are (a) profile plots of the E-residuals or first stage residuals; (b) Scatterplot matrices (SPM) of the random effects (RE's); and (c) Scatterplot matrices (SPM) of the E-residuals.

• Reading in the data.
  • SAS program to read the data.
  • Alternate SAS program to read the data.
• Model 1, RE = intercept, day, visit, (visit independent of others).
  fixed effects = intercept, day, group, visit.
  • SAS program to fit the model
  • Unedited SAS output
  • Profile residual plot (postscript)
  • RE and residuals SAS outputfor input to Xlisp.
  • Xlisp-Stat program to produce two residual plots.
  • Plot of RE's (postscript)
  • SPM of Residuals (postscript)
• Model 2, RE = intercept, day, visit, (visit independent of others)
  fixed effects = intercept, day, group, visit, day*visit.
  • SAS program to fit the model
  • Unedited SAS output
  • Profile residual plot (postscript)
  • RE and residuals SAS outputfor input to Xlisp.
  • Xlisp-Stat program to produce two residual plots.
  • Plot of RE's (postscript)
  • SPM of Residual (postscript)
• Model 3, RE = intercept, day, visit. (All correlated)
  fixed effects = intercept, day, group, visit.
  • SAS program to fit the model
  • Unedited SAS output
  • Profile residual plot (postscript)
  • RE and residuals SAS outputfor input to Xlisp.
  • Xlisp-Stat program to produce two residual plots.
  • Plot of RE's (postscript)
  • SPM of Residual (postscript)
• Model 4, RE = intercept, day, visit, (visit independent of others)
  fixed effects = intercept, day, group, visit, day*visit, day*group.
  • SAS program to fit the model
  • Unedited SAS output
  • Profile residual plot (postscript)
  • RE and residuals SAS outputfor input to Xlisp.
  • Xlisp-Stat program to produce two residual plots.
  • Plot of RE's (postscript)
  • SPM of Residuals (postscript)
• Model 5, same as 1, but with REML fit.
  • SAS program to fit the model
  • Unedited SAS output
  • Profile residual plot (postscript)
• Model 6, same as 2, but with REML fit.
  • SAS program to fit the model
  • Unedited SAS output
  • Profile residual plot (postscript)
• Model 7, same as 3, but with REML fit.
  • SAS program to fit the model
  • Unedited SAS output
  • Profile residual plot (postscript)
• Model 8, same as 4, but with REML fit.
  • SAS program to fit the model
  • Unedited SAS output
  • Profile residual plot (postscript)