Modelling the participation function with a one-parameter family of cubic splines

 

We suggest that a simple one-parameter family of cubic spline functions would serve quite adequately as models of the participation curve that is the key component of ITI calculations. This would remove the subjectivity associated with the use of two-parameter logistic functions, and would allow all states to use the same method for ITI calculations.

Cite

@techreport{ITIspline15,
  author      = {Ken J Harrison and Rob J Hyndman},
  institution = {Monash University Statistics \& Econometrics Consulting Service},
  keywords    = {misc},
  note        = {{R}eport for {ITI} technical group, Australian Conference of Tertiary Admission Centres},
  title       = {Modelling the participation function with a one-parameter family of cubic splines},
  url         = {http://robjhyndman.com/publications/cubicsplinesITI/},
  year        = {2015}
}