Week 12: More advanced Methods - Factor Analysis Flashcards

1
Q

Factor analysis -

A

involves grouping similar variables into dimension / factors

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

T/F Multiple variables can be grouped into a couple of factors and a couple of factors pre defined and the relationship to the multiple variables can be tested

A

True

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

T/F Factor analysis is quite different from any of the statistical procedures that use data for comparison or prediction

A

True

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

Factor analysis is most commonly used in what approach?

A

an exploratory approach to data analysis (Exploratory Factor Analysis; EFA)

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

Exploratory Factor analysis is good for what 2 things?

A
  1. good to examine the structure within a large number of variables
  2. good to explain the nature of their interrelationships
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Factor analysis can also be used in what way?

A

also a confirmatory manner to data analysis (Confirmatory Factor Analysis; CFA)

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

Confirmatory Factor Analysis is used in what two ways?

A
  1. used to verify the factor structure of a set of observed variables
  2. used to test the hypothesis that a relationship between observed variables and their underlying latent constructs exists
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

T/F Factor analysis is more controversial than other analytic methods leaving room for subjectivity and judgment

A

True

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

What method is crucial to provide structure to data sets with multiple observed variables?

A

Multivariate analysis

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

EFA is a factor analysis used to do what?

A
  1. explore the possible underlying factor structure of a set of observed variables without imposing a predefined structure of the outcome
  2. identify the underlying factor structure
  3. describe and identify the number of factors
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What are 3 goals of EFA?

A
  1. to determine the number of latent constructs underlying a set of variables
  2. to provide a means of explaining variation among variables using a few newly created factors
  3. to define the content of meaning of factors
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

3 Assumptions underlying EFA?

A
  1. continuous level of measurement with normal distribution
  2. sample size should be large enough more than 200 and 5 observations per variable
  3. correlation > 0.3 between the variables
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Limitations of EFA?

A
  1. variables could be sample specific, not generalizable
  2. non-normal distribution of data
  3. sample size larger than the required is desirable to accommodate the possible missing data
  4. No causal inferences can be made from correlations alone
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

CFA is a factor analysis used to do what?

A

test the hypothesis that there exists a relationship between the observed variables and their underlying latent constructs

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

CFA procedure:

A
  1. review the relevant theory and research literature to support model specification
  2. specify a model
  3. collect data
  4. assess model fit by
    - hypothesis testing
    - fit indices look up
    e. g.) CFI (Confirmatory Factor Index)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Limitations of CFA:

A
  • sample size required to be sufficiently large
  • multivariate normality
  • outliers
  • missing data
17
Q

What happens if an unacceptable model fit is found in CFA?

A

an EFA can be performed to identify the underlying factor structure, describe and identify the number of factors

18
Q

When developing factors, what is a factor loading?

A

the coefficient as a measure of the correlation between the individual variable and the overall factor

19
Q

Criteria for significance of factor loadings (FL) when developing factors:

A

FL > 0.3: minimum consideration
FL > 0.4: more important
FL > 0.5: practically significant

20
Q

How do you perform extraction of factors?

A

Pull out only the components using a cutoff point where value at least 1

21
Q

What is rotation of factors?

A

a process of developing a unique statistical solution so that each variable relates highly to only one factor

22
Q

After rotation of factors, what do you do?

A

Naming factors

23
Q

3 Limitations of factor analysis:

A
  1. Be cautious about how ‘factors’ are interpreted they are not real measurement entities only hypothetical statistical concepts giving factor a name doesn’t make it real
  2. Data may be organized differently by using different extraction or rotation methods these differences can alter a factor’s essential meaning
  3. The generated factors may be totally uninterpretable within the framework of the research question
24
Q

T/F The application of factor analysis involves grouping similar variables into factors

A

True

25
Q

What type of analysis is used to explore the possible underlying factor structure of a set of observed variables without imposing a predefined structure of the outcome?

A

Exploratory factor analysis (EFA)

26
Q

What type of analysis is used used to test the hypothesis that there exists a relationship between the observed variables and their underlying latent constructs?

A

Confirmatory factor analysis (CFA)

27
Q

What are the steps of of an exploratory factor analysis ?

A
  1. developing factors
  2. extracting factors
  3. rotating factors
  4. naming factors
28
Q

Limitations of factor analysis stem from what?

A

subjectivity and judgmental nature in decisions:

  1. factors are not real measurement entities only being hypothetical statistical concepts
  2. the resulting data structure is subject to different selection of extraction or rotation methods
  3. the generated factors may be totally uninterpretable within the framework of the research question