Factor analysis - week 6

?
What is a uni dimensional scale?
Only measures one thing
1 of 18
What is a multi dimensional scale?
Measures more than one things
2 of 18
What is factor analysis?
Form of advanced multivariate correlational stats- correlates everything- looks for patterns of covariation (clusters of items)
3 of 18
What does factor analysis try to do?
Tries to determine whether the variation in these correlated questions/variables can be explained by a smaller number of variables we cannot observe.
4 of 18
What is common variance?
The variance in each question that can be explained by a common factor - a range of questions might be getting at some part of the variable measured
5 of 18
What is unique variance?
The variance in question that is caused by the uniqueness of that question. Each question has a bit of uniqueness
6 of 18
What is error variance?
The bit of randomness in question responses due to extraneous variables (mood etc)
7 of 18
So what does factor analysis do with these variances?
Tries to work out what the common variance is between items- uses this to say whether there are unerlying latent variables that explain some of the variability in the scores for the different items
8 of 18
What is exploratory factor analysis?
When we don't know how many factors are amongst our variables or we have tried confirmatory factor analysis- so any indicator can be linked with any factor.
9 of 18
What is confirmatory factor analysis?
When we know the number of factors and how they are made-up -thus we use a set number of items which we expect to be associated with distinct factors. We then check to see that this is the case in our dataset
10 of 18
What is the difference between PCA and factor analysis ?
PCA assumes all variance is common variance when that isn't necessarily the case- PAF OR ML estimates common variance for each variable then tries to reduce this into distinct factors.
11 of 18
What 2 things do we look at before running a factor analysis?
a) eigenvalues ovr 1 and b) the scree plot
12 of 18
How do you examine a scree plot?
Find the point at which the shape of the curve changes direction and becomes horizontal - point of inflection (elbow point)
13 of 18
What is oblique factor rotation?
No constraint on independent factors (so factors correlate with each other) - most common is direct oblimin rotation
14 of 18
How to choose rotation?
Run oblique rotation (direct oblimin) and see if the factors seem to be highly correlated. If they are not we can re-run using an orthogonal rotation (varimax)
15 of 18
Why do we need to rotate?
imagine we had a graph which had 2 dimensions corresponding to our factors - the point on graph are items - we rotate the axe to make sureties going directly through them questions.
16 of 18
what is orthogonal rotation?
The axis stay perpendicular, meaning that the factors remain maximally (completely) unrelated to each other. Varimax = orthogonal rotation
17 of 18
What is oblique rotation?
Assumes thee factors can be correlated. This enables the rotation to maximally fit both factors. direct oblimin= oblique rotation
18 of 18

Other cards in this set

Card 2

Front

What is a multi dimensional scale?

Back

Measures more than one things

Card 3

Front

What is factor analysis?

Back

Preview of the front of card 3

Card 4

Front

What does factor analysis try to do?

Back

Preview of the front of card 4

Card 5

Front

What is common variance?

Back

Preview of the front of card 5
View more cards

Comments

No comments have yet been made

Similar Psychology resources:

See all Psychology resources »See all Visual System resources »