Multivariate Analysis Flashcards

1
Q

Multivariate Analysis

A
  • Statistical test that lookes at potential interrelationships in the data
  • Specifically, for more than one dependent variable at once
  • A new dependent variable is formed from a linear composite of the dependent variables
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

If there are multivariate differences…

A

further analysis is justified

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

If there are no multivariate difference…

A

analysis stops

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Why are multivariates useful?

A
  • Useful to examine one than one dependent variable, especially if they are related!
  • Thus we can look at how an individual varies on all of the parameters concurrently!
  • Helps decrease type 1 error/familywise error
  • More powerful (more likely to see difference) than ANOVA
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

How many subjects do you need per variable for multivariate analysis?

A

At least 10 per

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Multivariate Analysis must be ____ of each test

A
  • independent
  • Ex: Torque and Force are related. Can’t compare them
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Most often used thing we read on SPSS to know if significance is there?

A
  • Wilk’s Lambda
  • Is like the F or T value
  • If this is significant run a post-hoc!
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Assumptions for Multivariate

A
  • Use common sense if it should be used
  • Better to use when relationships among dependent variables are r=0.3-0.7
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Multivariate cautions

A
  • Ratio of subjects to dependent variables should be greater than 3:1 (Close 1:1 severly limits power)
  • Outliers severly affect MANOVA findings
  • Easy to violate assumptions (Homogenity of variance)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly