How to run the Univariate procedure

Prepair the Data

 

A. A 2 x 3 Factorial Design Example

Variable 1: Learner proficiency: high vs. low
Variable 2: Treatment type: A, B, C
Dependent variable: test scores

case L2prof treatmnt  scores
1 high A  67
2  high A  42
3  high A  45
4  high B  61
5  high B  64
6  high B  57
7  high C  43
8  high C  41
9  high C  56
10  low A  53
11  low A  79
12  low A  63
13  low B  58
14  low B  59
15  low B  61
16  low C  60
17  low C  51
18  low C  53

* In the data set for a univariate analysis, each independent variable takes a column, starting from the more general variable. This procedure is for a between-subject design, so each row represents a subject and the same subject can't appear in more than one row/cell.

B. A 2 x 2 x 2 Factorial Design Example

Variable 1: L2 proficiency: High vs. Low
Variable 2: word frequency: High vs. Low
Variable 3: word type: noun vs. verb
Dependent variable: mean reaction times

L2prof

 wordfreq wordtype meanRT
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L

HF
HF
HF
HF
HF
HF
HF
HF
LF
LF
LF
LF
LF
LF
LF
LF
HF
HF
HF
HF
HF
HF
HF
HF
LF
LF
LF
LF
LF
LF
LF
LF

N
N
N
N
V
V
V
V
N
N
N
N
V
V
V
V
N
N
N
N
V
V
V
V
N
N
N
N
V
V
V
V

603
523
529
543
876
638
980
653
1087
903
1120
1058
1382
1429
1235
1016
737
686
609
672
619
622
707
1632
732
1054
1639
1459
1465
1118
1638
1353

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2. Run the Procedure

 

1. Move the dependent variable into the Dependent Variable space and the independent variables into the Fixed Factor(s) space.

2. Click on OK to run the procedure.

3. A covariate is a variable that are not directly tested but may correlate with the dependent variable, e.g. IQ. If this variable is moved into the Covariate box, the effects of this variable ("data noise") are removed and thus the ANOVA is performed on a purified data set resulting in greater power. This technique is known as Analysis of Covariance (ANCOVA).

4. Click on the Post Hoc button to make pairwise comparisons (transfer varaible, check Tukey, then Continue, OK. Or run
a One-Way ANOVA with the Tukey test
for each independent variable.
 

Read the Output (Example B)

There is no main effect of L2 proficiency (F=2.20, p=.15); There is a main effect of word frequency (F=30.47, p <.01) and a main effect of word type (F=5.45, p=.028). There is no significant interaction.