Orthodont data set

Introduction to MLR with examples

Chapter 12 of a stats book for upper division applied stats course.

Analysis of a real data set on plant growth. Includes, notes on modeling, estimation and prediction. Created as a template to create a complete report for SLR.

A realistic distribution of number of insects per trap is manufactured using standard distributions. The Gini coefficient for the population is calculated and the sampling distribution of the estimated Gini coefficient is created by numerical simulation. The population is sampled and the Gini coefficient is estimated and CI constructed by nonparametric bootstrapping. Bootstrapped CI coverage is studied by repeated "real" resampling from the simulated and known population and repetition of the bootstrapping procedure. Coverage is low (90/95) for a sample size of 20 but reaches nominal values for sample size 30.

I explain a couple of ways in which models are parameterized in R and give a procedure to create vectors that will estimate and test any hypothesis reducible to linear combinations of estimated parameters.