L8 - Factor Analysis Flashcards

1
Q

Purpose of the Factor-analysis

A

Discover the factors that influence the co-variation among multiple observed variables; reduce large set of variables to smaller set of factors.

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

Two categories of approaches concerning factor analysis

A
  • factor analysis (estimation of latent variable and generalization to population)
  • exploratory PCA
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What are two types of factor analysis

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

What is PCA about?

A

Summary of correlational structure in a given data set

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

are people in the data set objects or items?

A

objects

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

How does the model look like in factor analysis?

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

What are the loadings?

A

weight of a factor to express x

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

How to determine how many factors to retain

A

Eigenvalues

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

What is an Eigenvalue

A

Sum of the (normalized) variance that is reproduced by a specific factor.

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

How to decide how many factors to retain?

A

Put eigenvalues on y-axis and components on x-axis:

  • Scree test
  • Parallel analysis
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is the scree test about?

A
  • Plot eigenvalues of factors in descending order
  • determine the number of factors where the eigenvalue levels off “Elbow criterion”
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is the parallel analysis about?

A
  • You generate a random data set with the same number of variables and objects as the empirical data set
  • run PCA
  • repeat many times
  • plot average eigenvalue for each extracted value

–> cutt off where data and simulated data meet

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

The Elbow criterion is used to

A

identify an appropriate number of clusters and factors

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is factor loading?

A

The correlation of an item with a factor

e.g item 21 with factor PC1

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is communality?

A

Proportion of variance of a variable accounted for by the extracted factors.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What is uniqueness

A

1 - communality. Proportion of variance of a variable not accounted for by the extracted factors.

17
Q

How can you express the correlation?

A

As an angle between vectors.

18
Q

What does an angle between vector of <90° indicate?

A

positive correlation

19
Q

What does an angle of 90° indicate?

A

No correlation.

20
Q

What does an angle of >90° indicate?

A

negative correlation

21
Q

What facilitates interpretation of factors?

A

Rotation.

Here F1 and F2 better represent the data. You have a more equal distribution of variance across the PCs.

22
Q

What is orthogonal rotation?

A

Rotated factors are uncorrelated.

23
Q

What is oblique rotation?

A