MAVOVA Flashcards

1
Q

what is a manova?

A

An anova with multiple DVs

Multivariate analysis of variance

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

What is the DV in manova?

A

2 or more DVs are all measures of the same construct (e.g. 3 memory tests)

Linear composite (LC) is the linear combination of multiple DVs

Different combination of DV for each effect in the design

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

When can manova be used?

A

Same as anova:

repeated-measures, between-groups, factorial designs (more than 1 IV). can include covariates (MANCOVA)

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

Central question in manova

A

Are there mean differences between groups on a combination of DVs?

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

manova and DA (discriminant (function) analysis)

A

Mathematically the same

manova is DA in reverse

manova- compare 2 or more groups on mulitple DVs

DA- same data, but use linear composites to PREDICT group membership

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

Why include multiple DVs?

A

Capture broader and richer information

Good IVs might effect DV in more than one way, so include several DVs

e.g. stress intervention (IV) might cases changes in different variables (the DVS- physiological, emotional, observational responses)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What does manova inferentially identify?

A

1) different levels of the IV have a significant effect on a liner combination of DVs
2) there are interactions between IVs and the LC
3) WHERE the effect is- are there significant univariate effects for each of the DVs?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

When does manova work well?

A

manova works well when there is MODERATE correlation between DVs

Too high (>.8)= not enough variance left after the first DV

Too low (

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Linear composites (LCs)

A

A linear combination of the DVs

Different to univariate approach where for 3 DVs, there would be 3 separate anovas

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

LCs in a factorial design

A

Different LCs are created for each main effect and interaction, same with comparisons

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

mavova F statistic

A

If the manova F statistic is significant, LC differs as a function the IV

:( difference in the LC could be due to IV differences among any, or all of, the DVs- it doesn’t tell you where the differences lie

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

How are DVs combined in the LC?

A

DVs are combined to separate the groups as much as possible

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