Introduction to the Design and Analysis of Experiments

Some basic statistics, especially hypothesis testing and inference for normal distributions, estimation and confidence intervals in linear models, linear regression; purposes of experimentation, sources of variability, randomization, replication, blocking, analysis of variance in simple designs. Methods of comparing treatment means and general contrasts among means. Latin squares and other orthogonal designs; assumptions behind an analysis. Study of plot residuals. Transformation of data; factorial sets of treatments. Factors at 2 and 3 levels. Interpreting interactions. Methods of analysis. Confounding to limit block size. Linear and quadratic contrasts for 3 levels; missing, lost or mixed-up observations. Methods for orthogonal designs. Iterative computing methods; analysis of convariance, including its uses as a missing-value technique; crossover designs. Analysis of medical trials; split-plot and split-block designs; non-orthogonal designs, total balanced oncomplete blocks, row and column designs, Youden squares and extensions; partial confounding and fractional replication in simple factorial designs; fitting response surfaces when quantitative factors are used, design of experiments for response-surface analysis; matrix notation for analysis of general block designs. Use of standard computer packages.