Lecture 7

Question 1

Factor analysis

Answer 1

Technique to reduce a large amount of information (contained in a number of original variables), into a simple message (fewer variables), with a minimum loss of information

Question 2

Latent variable

Answer 2

Thing that cannot be measured directly, or cannot be observed

Question 3

Factors (latent variables)

Answer 3

Clusters of large correlation coefficients suggest that those variables could be measuring aspects of the same underlying dimension.

Question 4

Factor loading

Answer 4

Correlation between a factor and a variable

Question 5

Extraction

Answer 5

Process of how many values to keep, by using eigenvalues

Question 6

Eigenvalues

Answer 6

Indicate the substantive importance of a factor. Retain only the factors with a large eigenvalue (>1)

Question 7

Scree plot

Answer 7

Graph eigenvalues (y axis) against the factor with which the eigenvalue is associated. Cut-off at the point of inflexion, You will retain 2 factors that are to the left of the reflexion point, you don’t include the inflexion point itself.

Question 8

Point of inflexion

Answer 8

Point at which the slope curve changes dramatically.

Question 9

Factor rotation

Answer 9

Visualise the factors as axes. This is done to make sure that the variables load maximally to one factor and have +/- 0 loading on the other factor.

Question 10

2 procedures of factor rotation

Answer 10

• Orthogonal - independent/ uncorrelated factors

- Obligue - allows factors to be correlated

Question 11

Reliability

Answer 11

Extent to which a variable is consistent in what it is intended to measure

Question 12

How to meaasure reliability

Answer 12

Cronbach’s alpha, it is required to be >0,8 or >0,7, otherwise, unreliable