Joint models for mean and dispersion

Code for analysis of robust design experiments with time dependent ordinal response characteristics in R

  • Example data from a horticultural quality improvement study
  • Code in file md models fits repeated measures proportional odds logistic regression model to time course of ordinal scores to give rate of quality decline for each plant and then fits linear models for mean and dispersion of rates for each plant
  • Model fitting requires installation of the R library gee and code for fitting repeated measures proportional odds logistic regression models is available using repolr