ST711 - Experimental Design
- Prerequisites: ST 512
- Term & Frequency: Fall, yearly
- Student Audience: Graduate students from any discipline wanting to design experiments.
- Credit: 3 credits
- Recent Texts: Planning, Construction, and Statistical Analysis of Comparative Experiments by Giesbrecht and Gumpertz.
- Recent Instructors: D. A. Dickey, J. A. Osborne, Jon Stallings
- Background and Goals: As background, students should understand completely randomized designs and the concept of blocking. An understanding of how linear regression works, especially in terms of matrix computations, is extremely helpful as background. Coming out of the course the student should understand the many different kinds of blocked designs, complete and incomplete, Latin squares, split plots, repeated measures, etc. (see list below) and the role of random effects in their analysis.
- Content: Review of Model Parameterizations, F-statistics and Randomization Tests, Nonparametric Tests, Power and Sample Size, Estimability of Effects, Blocking, REML Estimation and PROC MIXED, Latin Squares, Youden Squares, Split Plots, Repeated Measures and Crossover Designs, Incomplete Block Designs, Lattice Designs such as Alpha Arrays, Two Level Factorials, Experiments without Replication, Factorials in Incomplete Blocks, Confounding , Fractional Factorials, Response Surface Methodology
- Alternatives: None
- Subsequent Courses: ST 524, ST 641, ST 755
S1 2017 Sections:
|-||/ - |