The Winer Datasets

1. B.J. Winer,
   Statistical Principles in Experimental Design
   McGraw-Hill, New York, 1970
   2nd edition

   the classical Winer, which will in general no more available,

2. B.J. Winer, D.R. Brown, K.M.Michels
   Statistical Principles in Experimental Design
   McGraw-Hill, New York, 1991
   3rd edition

The Format

In case of repeated measurements (trial factors) the indices of the first trial factor are moving slowest, the indices of the last trial factor are moving fastest. This is the convention of most computer programs doing such analysis (e.g. BMDP, SPSS). In case of covariates these preceed the dependent variable.

Therefore the general sequence is:
levels of 1. grouping factor, levels of 2. grouping factor, ... 1. covariate, 2. covariate, ... dependent variable repeated for the levels of the trial factors.

The Data

Single Factor Designs

[2] page 115, table 3.13

Design:

1 grouping factor (treatment) with 4 levels
26 subjects (unbalanced)

[2] page 217, exercise 2

Design:

1 grouping factor with 4 levels
1 control group (4th level)
44 subjects (balanced)

[2] page 228, table 4.3

Design:

1 trial factor with 4 levels
5 subjects

[1] page 136, table 4.7-1

Design:

1 trial factor with 4 levels
8 subjects

[1] page 248, table 4.11

Design:

1 trial factor with 3 levels
5 subjects

Designs with several Grouping Factors

[2] page 425, table 6.3

Design:

2 grouping factors (drugs, schizophrenics/depressives) with 2 and 3 levels,
18 subjects (balanced)

[2] page 439, table 6.10

Design:

2 grouping factors with 2 and 4 levels,
33 subjects (unbalanced)

[2] page 468, table 6.25

Design:

3 grouping factors with 3 levels each,
27 subjects (one observation per cell)

Multifactor Repeated Measurement Designs

[2] page 518, table 7.5

Design:

1 grouping factor with 2 levels,
1 trial factor with 3 levels
10 subjects (balanced)

[2] page 521, table 7.7

Design:

1 grouping factor with 2 levels,
1 trial factor with 4 levels
6 subjects (balanced)

[2] page 537, table 7.13

Design:

1 grouping factor (noise) with 2 levels,
2 trial factors (periods, dials) both with 3 levels
6 subjects (balanced)

[2] page 552, table 7.22

Design:

2 grouping factors (anxiety, low/high) with 2 levels each,
1 trial factor (trials) with 4 levels
12 subjects (balanced)

[2] page 568, table 7.30

Design:

2 grouping factors with 2 levels each,
1 trial factor (trials) with 4 levels
12 subjects (balanced)

[2] page 578, table 7.38

Design:

1 grouping factor with 3 levels,
1 trial factor with 3 levels
12 subjects (unbalanced)

Designs having Covariates

[2] page 776, table 10.7

Design:

1 grouping factor with 3 levels,
1 covariate 21 subjects (balanced)

[2] page 787

Design:

1 grouping factor with 3 levels,
1 covariate (having 4 values to be used as a factor) 24 subjects (balanced)

[2] page 794, table 10.16

Design:

1 grouping factor with 2 levels,
2 covariates 24 subjects (balanced)

[2] page 811, table 10.22

Design:

2 grouping factors with 2 and 3 levels,
1 covariate 30 subjects (balanced)

[2] page 818, table 10.29

Design:

2 grouping factors with 2 and 3 levels,
1 covariate 31 subjects (unbalanced)

[2] page 829, table 10.34

Design:

1 grouping factor with 2 levels,
1 trial factor with 2 levels
1 covariate constant over trials 8 subjects (balanced)

[2] page 833, table 10.36

Design:

1 grouping factor with 3 levels,
1 trial factor with 2 levels
1 covariate changing over trials 9 subjects (balanced)

Special Designs

[2] page 633, table 8.25

Design:

2 trial factors (A, B) with 3 levels each, observed as
1 grouping factor: 6 blocks and
1 trial factor: 3 treatment combinations of A and B varying over blocks:
block treatment combination
1 (11) (23) (32)
2 (12) (21) (33)
3 (23) (22) (31)
4 (11) (22) (33)
5 (13) (21) (32)
6 (12) (23) (31)
12 subjects (balanced)