3B. Factor Analysis

Question 1

What is factor analysis primarily used for?

Answer 1

Data reduction and summarisation.

Question 2

Why do we need to reduce/summarise data using factor analysis?

Answer 2

• Customer surveys can be very long (often over 100 items)
• As marketers, we seek to derive insights from the survey. However, we don’t really care about a specific item, but rather what general factors drive customer behavior

Question 3

Factor analysis allows us to ________________

Answer 3

Factor analysis allows to narrow down a large number of specific observable (and often highly correlated) items to a few composite underlying factors

Question 4

What are the two types of “Factor analyses” in the module?

Answer 4

1. Exploratory Factor Analysis
2. Confirmatory Factor Analysis

Question 5

What is the difference between “exploratory factor analysis” and “confirmatory factor analysis”, and which do we use?

Answer 5

Exploratory = NO prior knowledge on how variables are related to each other

Confirmatory = there IS this knowledge..

== We use EXPLORATORY in this module

Question 6

From an “Exploratory Factor Analysis”, which are the two relevant sub-analyses?

Answer 6

1. Principal Component Analysis (PCA)
2. Common Factor Analysis (CFA)

Question 7

What is the difference between “Principal Component Analysis” and “Common Factor Analysis”?

Answer 7

BOTH are sub-analyses for “Exploratory Factor Analysis”..

BUT:
- PCA: considers total information (variance) in data
- CFA: considers only common variance of data

Question 8

Show the diagram of an “Exploratory Factor Analysis”?

Answer 8

• x is each variable (eg. willingness to pay, customer satisfaction), always denoted by x
• ε epsilon is the error term (per variable), error in the way we measure particular variables
• F is a factor, all items can correlate with every factor

Question 9

What is the goal of “Exploratory Factor Analysis”?

Answer 9

The goal of factor analysis is to group the variables that are most correlated to each other

== To divide variables into groups

Question 10

The aim of EFA is to separate the variables that __________

Answer 10

The aim of EFA is to separate the variables that correlate highly from those that correlate less strongly

Question 11

What is the difference between correlates within groups VS between group (EFA)

Answer 11

Within groups, variables should be as highly correlated as possible

… between groups, as low as possible!

Question 12

What is the factor in “Exploratory Factor Analysis”?

Answer 12

The factor can be seen as an underlying/ hidden variable
-> that affects several observed variables (the ones we measure)

// Several variables are observable phenomena of less underlying factors

Question 13

How does a Factor explain how observed variables are correlated?

Answer 13

Factor analysis groups/combines variables that are highly correlated with each other. It assumes those correlations happen because the variables share an underlying unmeasurable (latent) factor—a hidden construct that influences them (e.g., “anxiety,” “intelligence,” “satisfaction”).

–> THAT is the FACTOR

Question 14

How do you determine what a FACTOR means/ what you should name it?

Answer 14

In factor analysis, you don’t “discover” a factor’s name from the math!

You discover a pattern of shared variance, and then humans decide what it means and what to call it.

The researcher/analyst ultimately names the factor.

Question 15

What is an “EIGENVALUE”?

Answer 15

Total variance of all variables accounted for by ONE factor

Question 16

What is “FACTOR LOADINGS”?

Answer 16

Correlations between the variables and factors (ranges from -1 to +1)

ie: How much does one variable load into a factor

Question 17

What is “FACTOR SCORES”?

Answer 17

Relation between observations and factors

Question 18

What is “COMMUNALITY”?

Answer 18

Proportion of one variable’s variance explained by all factors extracted

ie: How good one factors solution is, how much of the info in each variable is explained by all factors, (eg. 80% of the variable is being explained by each factor)

Question 19

What is the OPPOSITE of “COMMUNALITY”?

Answer 19

EIGENVALUE

why?
They’re “opposites” in the sense that they answer two different directions of the same question:
- Communality: How much of this variable is explained by the factors?
- Eigenvalue: How much total variance is explained by this factor?

Question 20

Show the PROCESS chain for an EFA

Answer 20