Practical Data Analysis Examples

This directory contains examples from Brian Yandell's book Practical Data Analysis for Designed Experiments. File suffixes used in subdirectories include the following:

	*.dat		data
	*.sdat		data for S or Splus (if there are missing data)

	*.sas		SAS instructions (sometimes called "source")
	*.log		SAS log (check for errors)
	*.lst		SAS listing (sometimes called "output")
	*.prt		SAS condensed listings (from the saspr command)

	*.s		S (or Splus) instructions (or source files)
	*.sink		S listing
	*.script	S session using UNIX "script" command

	*.tex		LaTeX source
	*.ps		Postscript
	*.html		HyperText

Practical Data Analysis for Designed Experiments
Data Sets Listed by Book Parts: A|B|C|D|E|F|G|H|I
Alphabetical List of Data Sets

Proper credit: If you use an algorithm, dataset, or other information from here, please acknowledge both PDA and the original contributor of the material.

A. Placing Data in Context

Design

B. Working with Groups of Data

Tomato | Cloning | Diet

C. Sorting out Effects with Data

F. Regressing with Factors

G. Deciding on Fixed or Random Effects

Random | Select | Wheat

H. Nesting Experimental Units

I. Repeating Measures on the Same Subject

Real Data Sets

Alfalfa: analysis of covariance in nested design
Bacteria: balanced factorial; philosophical challenge in nesting: assignment of treatment to egg or room?
Berry: RCB, split plot, repeated measures
Brassica: RCB, 2 years, split plot, slight change year to year
Budding: 3 levels, additive and dominance?
Chew: two periods, polynomial contrasts
Cloning: one-factor ANOVA, multiple comparisons, unequal variance
Diet: analysis of covariance with blocks
Drink: complicated crossover design
Drying: nested design with covariate
Feed: one factor ancova with polynomial contrasts
Forage: split plot with two-factor in whole plot
Grain: RCB on farm trial
Growth: missing cell, transform, zeroes
Hardy: one factor with many levels, many responses; nesting; two-factor, ancova
Iris: ancova / discriminant analysis
Iron: error in variables
Oocyte: unbalanced factorial design
Paddock: complicated repeated measures
Potato: subsampling
Product: Taguchi L18 design -- planned missing cells
Rumen: split plot, replicated Latin square
Running: blocking, 3 factors
Season: repeated measures, missing data
Select: completely nested, fixed effects, selection
Tomato: nested design over 2 years, contrasts
Tree: repeated measures
Wasp: ancova / discriminant analysis
Wheat: completely nested, random effects

Artificial Data Sets

Biotron: RCBD
Bst: unbalanced factorial, contrasts
Company: RCBD
Damage: error in variables
Genss: simulate distribution of sums of squares
Leprosy: ancova from Snedecor and Cochran
Prefer: CRD and RCB (fictional data)
Random: random effects models
Strip: split strip plot (fictional data)
Tukey: Tukey and Mandel interaction plots/tests
Variance: testing with unequal variances

Other Schematic Plots

These do not have data associated with them.

Carry: cross-over design
Design: design schematics
Missing: missing data schematics
Nested: nested design schematics
Path: path diagrams
Random: random effects

Last modified: Mon Mar 2 14:34:14 1998 by Brian Yandell (yandell@stat.wisc.edu)