Random Coefficients Regression RPD – Section 18.3
Basic Model Simple Linear Regression where each of n experimental units is observed at t points in time (typically)
General Model (Gumpertz and Pantula (1989)) Possibly Multiple Linear Regression where each of n experimental units is observed at t points in time, based on regression with k parameters
Estimating Individual/Population Regression Parameters
Estimating Variance Parameters - I
Estimating Variance Parameters - II
Example – Annual Air Revenues for 10 Markets Random Sample of n = 10 large air markets (City Pairs), each observed over 5 years Y = ln(Average Fare * Average weekly Passengers) X = Year (1996/7=0, 2000/1=4) – Note: All Cities have same levels of X (not necessary for the method)
Air Revenue Data II