18 - Factor Analysis

Question 1

Factor analysis

Answer 1

Places together closely related items for a theoretical concept. It seeks to discover if the observed variables can be explained in terms of much smaller number of super-variables.
- EFA
- CFA
Factor loading should be at least 0.3

Question 2

Explanatory factor analysis

Answer 2

large data sets of multiple variables with smala factors thereby identifying factor structure

Question 3

Confirmatory factor analysis

Answer 3

Aims to confirm theoretical predictions = hypothesis testing.

Question 4

Principal component analysis

Answer 4

As there are an infinite nb if equally accurate factors, researcher requires a tool that chooses the most appropriate.

In EFA is used for data reduction or determining the set of items that hang together.

Question 5

PCA assumptions

Answer 5

• large sample (x5 or 10 variables)
• normal variables
• linear relation between variables
• absence of outliers
• interval data

Question 6

Communality

Answer 6

Sum of squared loading factors

Question 7

Eigenvalue

Answer 7

Illustrates the amount of common variance explained by a factor.
Sum of squared loading factors of the factor.

Question 8

Variance extracted

Answer 8

Eigenvalue*100%/nb ouf factors

Question 9

Factors issues

Answer 9

• naming

- selection of factors (kaiser’s rule >1 or scree test) not the only one possible

Question 10

Factor analysis steps

Answer 10

• descriptive and correlation matrix
• PCA gives an estimate of the loadings
• factors are rotated (varimax)
• interpretation